You see a dog in a park, standing near a bench while its person scrolls. The leash is loose. The dog is not performing, not begging, not doing anything dramatic. Then the person looks up, and the dog looks back, as if eye contact were a small private language.
It is tempting to turn that moment into a sweeping claim: dogs are built to read us. And maybe they are, in part. But psychology always has to ask the next question, the one that makes the claim less tidy and more useful: which dogs taught us that?
The Dogs Who Make It Into the Lab
For decades, many of the most elegant studies in dog psychology have depended on pet dogs. That makes sense. Pet dogs can be recruited through owners, transported to university labs, handled by strangers, vaccinated, filmed, and asked to tolerate strange rooms with quiet floors and unfamiliar smells.
These are not fake dogs. They are real dogs with real cognitive lives. But they are also dogs with a very particular biography.
They sleep in human homes. They hear their names. They learn the difference between a command, an apology, dinner gossip, and the sound of a cupboard opening. They discover that human attention can lead to food, doors, play, comfort, and praise.
So when a pet dog follows a pointing finger, gazes at a human face, or waits for permission, we may be seeing canine cognition. We may also be seeing years of close human tutoring.
A 2026 Psychological Science summary framed the concern plainly: canine science has been skewed toward the Western world, especially toward familiar companion-dog contexts and laboratory settings. The point is not that this research is worthless. It is that every sample carries a history. If that history disappears, the finding starts to look more universal than it may be.
Most Dogs Do Not Live on Sofas
The denominator matters. The World Organisation for Animal Health estimates that there are more than 700 million dogs globally, and around 75 percent roam freely. Many are street dogs, village dogs, campus dogs, market dogs, loosely owned dogs, or dogs whose lives are threaded through human spaces without being organized around a living room.
That changes the question. If most dogs are not Western pets, then the average laboratory dog may be ordinary in one setting and unusual globally. Both can be true.
A 2024 scoping review of dog behavioral testing found 392 publications and 2,362 behavioral tests, with striking variation in methods. Variation is not automatically a flaw. Different questions need different tools. But it should make us cautious when a headline says dogs look to humans for help, or dogs understand pointing, as if the sentence had no footnote.
The missing footnote is the real story: which dogs, raised where, tested by whom, rewarded how, and under what pressure?
Same Eyes, Different Social Economies
Brubaker and colleagues offered one useful comparison in 2019. They studied free-ranging dogs, shelter dogs, and pet dogs, asking how a human’s attentional state changed the dogs’ gazing behavior.
The free-ranging dogs reduced looking when the human became inattentive. Pet and shelter dogs gazed more overall and stayed closer to the experimenter. Nothing about that result turns one group into a better or worse dog. It suggests that attention has an ecology.
A pet dog may learn that staring at a person works. A street dog may learn that prolonged staring is risky, useless, or too costly. Same eyes. Different social economy.
Pointing shows the same complexity. Many pet-dog studies have found impressive sensitivity to human gestures. But in India, Bhattacharjee and colleagues tested free-ranging dogs in public spaces: 83 adults, 58 juveniles, and 68 pups. Pups followed nearby pointing readily. Adults adjusted after reward experience. The finding was not simple incapacity. It was plasticity shaped by age, learning, and street life.
That is one of psychology’s hardest lessons: a failed task can mean many things. No ability, no motivation, fear, unfamiliar setting, poor reward, sensible caution. A task is never just a task. It is a conversation between the method and the participant’s history.
The WEIRD Problem Has a Canine Cousin
This is where dog psychology becomes a doorway into psychology more broadly.
In 2010, Joseph Henrich, Steven Heine, and Ara Norenzayan argued that psychology relied too heavily on WEIRD participants: Western, educated, industrialized, rich, democratic samples. Their point was not that Western students are invalid humans. It was that they are a narrow slice of humanity.
Dog science risks a canine version of the same move. Western pet dogs become dogs. Other dogs become exceptions.
Animal behavior researchers Webster and Rutz have a useful term for this: STRANGE samples. The acronym asks researchers to consider social background, trappability, rearing history, acclimation, natural responsiveness, genetics, and experience. In plain language: who is able to show up for science?
For dogs, that often means an owner with time, transport, trust in a university, and a dog comfortable enough to enter the testing situation. That does not make the findings fake. It makes them situated.
Even human interpretation is sampled. In 2019, Amici and colleagues tested 89 adults and 77 children on recognizing dog emotions across cultural backgrounds. Anger and happiness were recognized broadly, but adults raised in dog-positive cultural settings read dog emotions better overall. A dog’s face meets a human history before it becomes obvious.
The More Humane Question
The answer is not to discard pet-dog research. Pet dogs are evidence. They are just not automatic evidence for every dog everywhere.
Free-ranging dogs are not a pure alternative either. A market dog, campus dog, and village dog may live very different lives. The goal is not to replace one stereotype with another. It is to keep the sampling clause attached to the claim.
The ManyDogs Project is one hopeful response: a collaborative attempt to study dog cognition across multiple sites, methods, and populations with more transparency.
So the next time a study says what dogs are like, pause before believing or dismissing it. Add one sentence first: who was actually studied, and whose life was left out?
That question may be one of the most humane tools psychology has.
