Can Dogs Smell Fear? What the Research Really Shows

By Dr Duncan Houston

“Dogs can smell fear” can sound unsettling if you are already nervous around them. The research behind that phrase is interesting, but it does not mean a dog can read your thoughts or that feeling afraid makes you responsible for its behaviour.

Smell is one source of information in an encounter. The surroundings, the dog's experience, your movements and the space available also matter.

Quick Answer

Dogs can detect differences between human odour samples collected under some stressed and unstressed conditions. Experiments also suggest that these odours can influence dogs' behaviour. This is evidence about chemical cues in controlled settings, not proof that every dog identifies every emotion or reacts to a nervous person in the same way.

What does “smelling fear” actually mean?

Researchers collect body-odour samples while people take part in particular tasks or emotional situations. Dogs are then exposed to those samples under controlled conditions.

That allows researchers to ask whether smell alone carries information, without the donor's face or voice being present. It does not turn fear into one universal smell that every dog interprets identically.

Keep three questions separate:

• Detection: can the dog tell two samples apart?
• Response: does exposure change the dog's measured behaviour?
• Interpretation: what, if anything, does the dog understand about the person's experience?

An experiment may answer the first question without answering the third. Saying a dog distinguishes an odour is more precise than saying it knows why a person is frightened.

Three studies, three different questions

Fear-associated sweat changed behaviour in retrievers

In D'Aniello and colleagues' study, published online in 2017 and in Animal Cognition in 2018, Labrador and Golden retrievers encountered samples collected from people in fear or happiness conditions, or a no-sweat control.

The dogs could move around a room containing their owner, an unfamiliar person and an odour dispenser. In the fear-odour condition, researchers recorded more stress-related behaviour and higher heart rates than in the other conditions. This supports an effect of the experimental odour exposure; it does not establish that a dog approaching you outdoors will be protective or aggressive.

Trained dogs distinguished samples before and after a stressful task

In Wilson and colleagues' 2022 experiment, four dogs learned a sample-choice task using combined human breath and sweat. They distinguished samples taken after a mental-arithmetic stress task from the same person's baseline sample and a blank.

The combined accuracy was 93.75% across the test trials. That figure belongs to these trained dogs and this test design. It is not the probability that a pet correctly identifies a stranger's feelings. The experiment demonstrated discrimination, rather than showing which individual molecule the dogs used or what meaning they assigned to it.

Human odour influenced one type of food-bowl choice

A 2024 study by Parr-Cortes and colleagues tested 18 dogs using odours from three unfamiliar human donors. The dogs learned which bowl locations contained food, then encountered bowls at uncertain locations.

Stress odour was associated with a reduced approach to one ambiguous location in the third session compared with the no-odour condition. The researchers interpreted this as a possible more cautious or “pessimistic” response. The effect depended on the location and testing session; it was not a general finding that every dog becomes unhappy whenever a person is stressed.

What the evidence does not tell you

These studies do not provide a way to predict an individual dog's intentions from your own feelings. They also do not establish that a dog smells a single “fear hormone” in every encounter.

Body-odour samples contain mixtures of chemicals. Research that shows discrimination between samples does not, by itself, identify the exact chemical cue. Claims that a dog is definitely detecting your cortisol or adrenaline need more specific evidence than a general scent experiment.

Outside a laboratory, several things change together. A nervous person might pause, become still, move differently or speak differently, while the dog is also responding to the location and other animals. One encounter cannot isolate smell from all those factors.

If you feel nervous around an unfamiliar dog

Make the interaction easier rather than trying to hide every sign of nervousness. Ask the handler to give you space, and do not feel obliged to greet or stroke the dog.

• Speak to the handler before approaching.
• Avoid crowding the dog or blocking its route away.
• Do not put your face close to its face or reach over it to prove you are friendly.
• If either of you is uncomfortable, end the proposed greeting and let the handler manage the dog.

The RSPCA's dog body-language guidance helps you look at the whole animal. A loose posture is different from marked stiffness, growling or attempts to move away. A wagging tail alone is not a guarantee that contact is welcome.

These are ordinary interaction precautions, not a test of whether the dog can smell your emotions. Choosing not to interact is a sensible option.

How owners can support a calmer situation

Give your dog predictable choices: enough distance, a quiet resting place and a familiar way to move away with you. Practise useful cues in an easy setting before expecting them to work during a difficult encounter.

Reward the behaviour you want, such as turning towards you or walking away calmly. The American Veterinary Society of Animal Behavior recommends reward-based training and arranging the environment so the dog can succeed. Forcing prolonged contact or punishing warning signals is not a useful way to test the scent hypothesis.

If a pattern keeps occurring, record the setting, what happened just beforehand, the distance involved and what helped the dog disengage. Avoid writing “smelled fear” as though it were an observed fact. “A visitor arrived; the dog became still beside the doorway” is a more useful description to discuss with an appropriate behaviour professional.

Common misunderstandings

• Assuming scent detection is the same as understanding a person's thoughts.
• Treating a laboratory accuracy figure as a rating for every dog.
• Calling every alert or close-following response “protection”.
• Blaming a nervous person for a dog's reaction.
• Trying to mask body odour instead of improving the situation.

Frequently Asked Questions

Will a dog attack because I am afraid?

The studies described here do not establish that prediction. Focus on the dog's behaviour, its handler's guidance and maintaining appropriate space.

Can my dog tell that I am stressed without seeing me?

Controlled experiments show that some human odour differences are detectable without the donor being visible. They do not tell you exactly what your own dog concludes from them.

Should I use perfume to hide fear from a dog?

There is no evidence-based reason in these studies to use fragrance as an interaction strategy. Ask for space and a manageable introduction instead.

Does my dog's closeness mean it is trying to protect me?

Closeness alone cannot establish the dog's intention. Observe the whole situation and avoid assigning a single explanation without supporting evidence.

A useful ability, with important limits

Dogs can gather information from odours we barely notice. The practical response is to respect that sensory world while arranging comfortable encounters, rather than expecting either person or dog to read the other's mind.

For more everyday pet-behaviour reading, explore ASK A VET™.