What Dogs May Notice About the People Around You

Author

Luboslav

Published

October 5, 2026

Reading time

5 min read

Topics

ContextObservation Method

A visitor enters, and your dog hangs back. Someone else receives an enthusiastic greeting. It is tempting to turn those reactions into a verdict on the people involved. Yet 'my dog avoided this person' and 'my dog knows this person is untrustworthy' are different claims. The first describes an event. The second assumes an ability that needs evidence.

Researchers have asked a narrower question: can dogs use an interaction they observe between humans when deciding whom to approach? This is called third-party social evaluation. It concerns particular actions under controlled conditions, rather than a general assessment of someone's character. Two studies show why the distinction matters.

▶Play video

A refusal to help changed one food choice

Chijiiwa and colleagues' 2015 study involved an owner trying to open a container holding a neutral object. An actor either helped after being asked, refused by turning away, or turned away without a request in a control condition. Dogs subsequently avoided accepting food from the person who had refused to help their owner. They did not show the same avoidance of the control actor. The result suggests sensitivity to that negative interaction; it does not establish a general preference for helpful people or an ability to detect bad character. [S1]

Notice what makes this task specific: the dog watched its own owner, the human action was a staged refusal, and the outcome was a food choice. A dog's response to a delivery driver, a nervous guest, or someone wearing unfamiliar clothing is not an equivalent test. Applying the result to those situations would require additional evidence.

A later experiment exposed a location problem

Jim, Marshall-Pescini and Range tested 42 dogs in 2020 using interactions between unfamiliar humans in a helping task. Some people remained in the places where the dogs had observed them; others switched sides. Dogs initially looked towards the helpful person only when the locations stayed the same. Their choices between people were random in both arrangements. The authors considered attraction to a location a possible explanation for the looking result. The study did not demonstrate robust reputation-based choice, but its specific task cannot rule out every form of social evaluation. [S2]

Together, these experiments leave an interesting question open. Observed conduct may influence dogs under some conditions, while place, familiarity and the available response can affect what a test reveals. A look towards someone, taking their food, and choosing to approach them should not be treated as interchangeable measures.

Your visitor brings more than a personality

Consider a guest who leans over your dog, speaks loudly and immediately reaches for its head. Compare that visit with someone who sits down and leaves the dog alone. Even if your dog approaches only the second guest, the difference is not a clean character comparison. The interaction styles changed. So did the distance, movement and opportunity to choose contact.

Now imagine that the quieter guest also sits beside the cupboard where treats are kept. That adds another possible influence. If the dog waits beside that chair on the next visit, you would need more context before attributing the behaviour to an opinion about the guest. This household example illustrates why location controls are useful; it is not a claim that all social preferences are simply food expectations.

The same caution applies when someone feels rejected by a dog. One hesitant greeting is limited information about the dog, the setting and the interaction. It need not become a judgment about the visitor, or a claim that the dog has failed to recognise a kind person.

Offer a greeting with room for choice

A sensible first visit gives the dog space and avoids making approach compulsory. Ask the visitor to settle before attempting contact. Provide an unobstructed route away and a resting place the visitor will leave alone. If your dog approaches, keep the interaction brief enough that it can end without anyone pursuing the dog. These are practical arrangements for a comfortable encounter, not techniques for proving a social-cognition theory.

Food should not become a reason to pull a reluctant dog into touching distance. If treats are involved, keep the setup spacious and avoid creating a conflict between wanting the food and avoiding contact. A visitor can remain welcome while the dog remains at a distance; successful hosting does not require an immediate friendship.

Record a reaction, not a verdict

Useful notes are concrete: who moved first, where the dog stood, whether a hand reached out, and whether the dog had room to retreat. Across ordinary visits, you may discover a repeatable pattern associated with an action or setting. That pattern gives you something specific to change.

A dog's preferences can matter without becoming evidence about a person's moral worth. Take discomfort seriously, reduce unnecessary pressure, and base decisions about people on human evidence. The research invites careful observation of interactions; it does not make a dog's greeting a character reference.

For a concise way to record your own observations, use the Dog Behavior Context Checklist.

Sources

[S1] Chijiiwa, H., Kuroshima, H., Hori, Y., Anderson, J. R., & Fujita, K. (2015). Dogs avoid people who behave negatively to their owner: third-party affective evaluation. Animal Behaviour, 106, 123–127. DOI: 10.1016/j.anbehav.2015.05.018.

[S2] Jim, H. L., Marshall-Pescini, S., & Range, F. (2020). Do dogs eavesdrop on human interactions in a helping situation? PLOS ONE, 15(8), e0237373. DOI: 10.1371/journal.pone.0237373.