Molinaro & Wynne (2025) — Anthrozoos
Barking Up the Wrong Tree: Human Perception of Dog Emotions Is Influenced by Extraneous Factors
Published: July 13, 2026
Most dog owners believe they can read their dog’s emotional state accurately. A two-experiment study involving nearly 900 participants has produced evidence that challenges that confidence directly — and the mechanism it identifies is one that affects every owner who has ever watched their dog in a specific situation and drawn a conclusion about how the dog was feeling. 🐾
Researchers H.G. Molinaro and Clive Wynne video recorded a dog in situations designed to be putatively positive or negative, then systematically manipulated what participants saw around the dog to test whether the surrounding context influenced emotional perception independently of the dog’s actual behaviour. The answer was unambiguous — and the degree to which context overrode the dog itself was larger than most participants would have predicted about their own judgements.
What Both Experiments Found
In the first experiment, 383 undergraduate participants were shown the same dog videos either with or without the visual background context. The presence or absence of context significantly affected how they rated the dog’s emotional valence — how positive or negative the dog’s emotional state appeared. Videos without context were rated as more positive than videos with context, regardless of what the dog was actually doing. Arousal ratings shifted as well, with positive situation videos rated as higher arousal specifically when context was removed. The background was not a neutral frame. It was an active variable in how the dog’s internal state was being read.
The second experiment pushed this further with 485 participants. Videos were deliberately edited so the dog appeared in mismatched contexts — placed against backgrounds that did not correspond to the original recording situation. The result was striking: when the human visible in the video was doing something positive, participants rated the dog’s emotional valence as higher and arousal as lower regardless of the actual situation the dog had been recorded in. The human’s behaviour in the background determined how the dog’s emotional state was perceived, even when the dog’s own behaviour told a different story.
Together the two experiments establish the same conclusion from two directions: what surrounds the dog in the scene is a major contributing influence on how humans perceive the dog’s emotions. Not a minor bias. A major one.
Why This Is a Problem for Animal Welfare and Everyday Ownership ⚠️
The welfare implications are direct. An owner who is smiling, relaxed, and engaged in something positive may consistently rate their dog as calm and content regardless of what the dog’s own posture, expression, and behaviour are communicating. A dog who is mildly anxious in a situation the owner reads as positive will have that anxiety consistently underread — not because the owner is inattentive, but because the human brain is assembling emotional perception from the entire scene rather than from the animal within it.
This effect is not limited to inexperienced owners or casual observers. The participants in this study were shown the same dog in the same conditions, and their judgements shifted systematically based on contextual variables that had nothing to do with the dog. The bias appears to be a feature of how human emotion perception works generally, applied to animals — and it operates below the level of conscious awareness.
The practical consequence is that any moment in which the surrounding context is emotionally loaded — a busy, positive social setting, a celebration, a training session going well — creates exactly the conditions under which a dog’s genuine distress or discomfort is most likely to be invisible to the people around them. The noise of the background drowns out the signal from the dog.
Reading the Dog Not the Scene 🐕
This is one of the most precise scientific descriptions of what NeuroBond works against. Real attunement is the deliberate practice of reading the actual animal in front of you — not the situation, not the mood of the room, not what you expect the dog to be feeling because of what is happening around them. It requires actively overriding the contextual pull that this research demonstrates is operating constantly and powerfully on human perception.
A dog at a family gathering who is showing subtle stress signals — ears back, low tail, body turned slightly away, reduced engagement — may be consistently perceived as fine because the surrounding humans are having a good time. The Invisible Leash functions only when you are actually receiving the signal the dog is sending rather than the signal the environment suggests they should be sending. The research is now quantifying exactly how difficult that distinction is to make — and how often the environment wins. 🐾
Source: Molinaro, H. G., & Wynne, C. D. L. (2025). Barking Up the Wrong Tree: Human Perception of Dog Emotions Is Influenced by Extraneous Factors. Anthrozoos. Published March 4, 2025.
