Do dogs have personalities?

By expanding our knowledge, we could get a better idea about matching dogs and people
By Sophia Yin DVM, November 2008, Updated June 2020
Photo by Jacob Curtis on Unsplash

Photo by Jacob Curtis on Unsplash

Every dog trainer or animal behavior consultant has probably heard this statement a million times: “My dogs Max and Emma are so different. Max comes when called and wants to please me, but if Emma is having fun, she just ignores me. They each have their own personality. ” While such references to personality in pets might rub the skeptical human behavioral scientist the wrong way, to animal behavior scientists—as well as anyone living with a dog or cat— it seems absurd to deny that personality exists in pets. In fact, using the same techniques that are used in human psychology, scientists have established that personality exists in a variety of animals—from monkeys and hyenas to zebra finches, garter snakes and octopuses.

Dr. John Capitanio, associate director for research for the California National Primate Research Center, explains this growing trend of studying personality in animals. “Personality is defined as consistent differences in behavior across time and context. There’s no reason to believe that animals such as rhesus monkeys shouldn’t be similar in terms of consistent personality traits, the way humans are. ”In a study Capitanio published in the American Journal of Primatology in 1999, he found that rhesus monkey personality could be described using four dimensions (sociability, confidence, excitability and equability) and that these dimensions correlated closely enough with behaviors in some situations that they could be used to predict behavior in other situations.

Temperament vs. Personality

Even among researchers, acknowledging consistent differences in behavior between individuals is nothing new, especially in companion animals such as dogs. For years, various groups and individuals have been developing canine “temperament tests.” Dr. Sam Gosling, a human-personality researcher and associate professor of psychology at the University of Texas, Austin, explains the terminology. “In human psychology, personality is different from temperament. Temperament is the early appearing, genetic-like traits. It is what you test in babies and young children. The idea is that temperament interacts with environment to comprise personality. With animals, that means that they have an innate temperament that interacts with environment to create personality.”

People who study animals tend to use the term “temperament” when what they really may be talking about is personality. Gosling, who is playing a pivotal role in expanding the study of personality to animals, says that while psychologists may see the application of human terminology as anthropomorphic, it may be more appropriate. “By failing [to use the term ‘personality’], we could needlessly be eliminating the idea that adult personality in animals is shaped by experiences.”


Sign up and get the answers to your questions.

Email Address:

Another reason to update the terminology is that it has an influence on the way research is performed; there is a long, rich history and methodology for studying personality in humans that can be applied to the study of animals. As the body of research expands, these studies can provide insight into not only animal behavior, but human behavior as well. For instance, while it can be difficult to follow people throughout their life to determine whether those with a certain personality will perform better or worse in specific contexts or environments, predictions about outcomes can often be tested over shorter time periods in animals.

When studying personality in animals, it is useful to first understand how personality is described in humans. The most firmly established system, devised from several decades of study, is the Five Factor Model (FFM). The five broad dimensions of personality in this model are openness, conscientiousness, extraversion, agreeableness and neuroticism. Each dimension includes clusters ofrecognizable traits, which are further subdivided into a number of descriptors.

For instance, as Gosling explains in his book Snoop: What Your Stuff Says About You, people who are rated as “open” tend to exhibit such traits as imagination, artistic interest, emotionality (awareness of their feelings), adventurousness (eager to try new activities and different experiences) and intellect (receptive to new and unusual ideas and eager to debate intellectual concepts).

Those scoring high in conscientiousness tend to be orderly, dutiful (have a strong sense ofmoral obligation), achievementstriving (focused), self-disciplined (persisting at difficult or unpleasant tasks until completed) and cautious.

People registering most strongly in the extraversion dimension are talkative and likely to be energetic (living fast-paced, busy lives), gregarious (enjoying the company of others and the excitement of crowds) and cheerful.

“Agreeable” individuals tend to be helpful, sympathetic, kind, forgiving and cooperative, and rate high in the facets of trust, morality (a low need for manipulation when dealing with others), altruism, modesty, cooperation and sympathy.

Finally, those who score high in neuroticism are easily stressed and anxious in tense situations. They tend to rate high in anxiety, anger (feeling they’re being treated unfairly), depression (lacking energy and having difficulty initiating activities), self-consciousness (easily embarrassed or ashamed) and vulnerability (panic, confusion and helplessness when under pressure or stress).

Based on this FFM model, it’s probably clear that personality in dogs and other animals is not completely analogous to personality in humans. Says Gosling, “In humans, it’s a much broader topic. Humans have traits as well as values and goals and identities and roles. When we talk about animal personality, we’re talking about a subset of what we call personality in humans.” For example, there is a discrepancy in the way that words are applied to humans and non-humans. In their 1997 study on personality in chimpanzees, researchers James King and Aurelio Jose Figueredo state that they used the term “lazy” to describe chimpanzees who were lethargic. This is different from the way that word is used to describe humans —a lazy person is often considered morally weak.

A 2008 study on personality in dogs done by Jacqueline Ley, Paula Bennett and Grahame Coleman of the department of psychology at Australia’s Monash University found a personality dimension in domestic canines that was similar to agreeableness but lacked the altruism qualities found in the human version of that dimension.

Words and Terms

Furthermore, some terminology used in human psychology research differs from that which has already been established in animal behavior research. For instance, “dominant” in human personality research describes the disposition of an individual, while dominance in animal behavior refers more precisely to status within a group. According to the American Veterinary Society of Animal Behavior’s position statement, The Use of Dominance Theory in Behavior Modification, dominance in animals is defined as “a relationship between individual animals that is established by force/aggression and submission to determine who has priority access to multiple resources such as food, preferred resting spots and mates.”

Capitanio, who studies monkey social structure, agrees. “Dominance is about social power, and as such, requires two individuals. Take the alpha animal from each of four different groups and put them together. You will not have four dominant animals! Rather, you will have a rank order of 1, 2, 3, 4, and only one dominant animal. Can you still describe the other three as ‘dominant’?”

Although human personality researchers may sometimes sneak human-specific terminology into their animal-related studies, those who study animal social structure or work in behavior modification maintain that dominance is not a personality trait in animals. Even so, traits such as aggression and confidence (or some combination of traits) that underlie personality may increase the likelihood that an animal attains a high rank and thus becomes dominant over other individuals.

These differences between humans and animals demonstrate the importance of species-specific research using the same techniques employed in human personality research. And herein lies the problem.

Early in this decade, Gosling and his graduate student, Amanda Jones, reviewed the state of personality research in canines. Their goal was to identify trends and any major shortcomings or unaddressed questions. A thorough literature search of all published research on the topic of temperament and personality in dogs yielded 51 studies.

Looking more closely at those studies, they found large gaps. For instance, the majority involved German Shepherd and Labrador Retriever puppies and were made for the purpose of predicting adult police dog or guide dog performance. Because substantial differences have been found in temperament among breeds, Jones and Gosling recommended that the findings of these studies not be generalized to the entire dog population. Age was another factor; as noted, most studies involved young dogs or puppies, and few included dogs over four years old.

Gosling and Jones felt that, since a primary use of personality studies might be to match shelter dogs with prospective owners, and such dogs are often over the age of four, age effects should be explored. Gosling and Jones found other problems with the body of research. There was no consistent use of terminology when describing temperament, which made it difficult to compare studies or to build upon other people’s work. And surprisingly few studies verified that independent raters used similar rating systems;thus, the reliability of most studies was questionable. Further, few studies tested the validity of the results or their ability to predict behavior.

Need for Consistency

For animal personality research to meet the standards used with humans, Gosling has proposed that the same criteria for testing personality in humans be met when testing to establish the existence ofpersonality traits in animals. For example, animals must be assessed by multiple independent observers, and their ratings must agree with one another more than 90 percent of the time—this is a test of interobserver reliability.

For reliability to exist, researchers need to identify parameters that observers can agree upon and choose appropriate people to rate the animals. In the case of dogs, “appropriate” people would be those who know them well—the people with whom they live. Shelter workers, whose contact with dogs is different both in context and duration, would not be appropriate to rate an animal’s personality in a shelter situation. Another impediment to accurate assessment in a shelter is that the setting is inherently stressful. As Gosling asks rhetorically, “Would you want a person to make a reliable assessment based on how you act in prison?”

Says Gosling, “When we say someone is friendly, they are friendly today and tomorrow and next year, in many situations. Evaluation based on one meeting would be like havingsomeone evaluate your punctuality based on whether you showed up on time [on one occasion]. Maybe you did, or didn’t. But there are many reasons other than personality why you would be late one time. It’s only after 20 meetings that you could say something about punctuality. Any single measure of behavior is unreliable.”

According to Gosling, a second criterion is that assessments must predict behaviors and real-world outcomes, meaning that the results can be proven to be valid and that they measure what we think they are measuring. Currently, there are no published and validated shelter temperament tests, which means that the tests that are being used may not predict actual future behavior outside the shelter. He points out, however, that researchers are looking at the various tests used in shelters to see if they yield similar results. For instance, if one test says the animal is fearful or nervous, does another test yield the same result? As Gosling notes, “If the results between the different tests are reliable, then the next step is to follow the animals home and see if the tests are valid—is the animal then [fearful or nervous] in the home?”

A third criterion for studying personality is that the tests should reflect an animal’s genuine attributes. For instance, one parameter used to describe an animal is openness, and a subtrait is curiosity. To prove that adjectives used to describe the animal are more than just projections of human ideas, one must perform behavioral tests. To test curiosity, for example, the researcher could test whether an animal investigates a novel object, or spends a significant amount of time exploring in a new environment. This way, the subjective rating categories are backed by behaviorbased factors rather than just words, which are more slippery and slide easily into anthropomorphizing.

Formulating a Method

With these criteria in mind, Jones began the meticulous processof developing a test that would evaluate personality in dogs. “First, we went through personality tests used in humans and compiled a list of 1,000 behavioral descriptors,” says Jones. “Then we got rid of the ones that were redundant. Next, we had a panel of seven dog experts look at the remaining list to see which were appropriate for dogs. This left us with 350 descriptors for our first questionnaire.”

Despite the great length of this questionnaire, they were able to find 3,000 dog owners who agreed to fill it out, determine whether items were difficult or easy to understand, and evaluate whether there were aspects the questionnaire did not cover. Based on the results of the initial test, they reduced the number ofquestions to 75 and then asked a new pool ofowners to fill out the shorter version.

Next, they performed a statistical analysis that looked for factors that could be grouped into broader dimensions. Says Jones, “We found four to five factors that describe a dog’s personality, depending on whether you divide aggression into that directed to owners versus that directed to dogs.” The dimensions they identified were fearfulness, aggression toward people, activity/excitability, responsiveness to training and aggression toward dogs.

Dogs who rated high in fearfulness scored high on questions related to whether they were nervous when greeting people, shy, easily startled, and fearful at the veterinary hospital or when restrained or groomed.

A high rating in the aggression-toward-people category reflected negative responses in the areas offriendliness around unfamiliar people, to a person approaching the house and when nervous or fearful, and positive responses to questions about aggression if disturbed when resting, when restrained, at veterinary visits and in connection with coveted items.

Dogs who rated high in activity/excitability were described as constantly active and/or boisterous, playful (enjoys playing with toys, tends to retrieve objects, can play for long periods of time), curious and actively engaged in tasks such as working to get treats out of a Kong toy. They also tended to seek out people and were more likely to be affectionate.

Those who rated high in responsiveness to training also ranked high in trainability (focused on a task, attentive to commands, responsive to corrections) and controllability (likely to come immediately when called, slow to sneak out through open doors and gates, quick to respond to corrections, likely to leave food or objects alone when told to do so).

Dogs were rated high in aggression to other animals if they tended to be aggressive to unfamiliar dogs, chase bikes or joggers, be excitable or aggressive toward cats, and guard food or treats from other dogs.

Even after paring down the questions to 75, Jones found that there were still redundancies. She also found that sometimes, just asking whether an animal was shy was good enough, and they could eliminate some of the questions pertaining to behaviors or body language indicating that the dog was shy or fearful. Says Jones, “It turns out that people can tell when their dog has a fearful personality, although they may not be able to make the connection that their dog’s aggression is due to fear.”

These categories were similar in some respects to those found in the 2008 personality study done by Ley and her colleagues at Australia’s Monash University. In their series of experiments, owners were given a list of 67 adjectives against which to rate their dogs. The Australian researchers found that adjectives tended to group into the following dimensions:extraversion, neuroticism, self-assuredness/motivation, training focus and amicability. Excitable, active, exuberant hyperactive dogs were considered to be extraverted; this is similar to the dimension Jones termed activity/excitability. Dogs who rated high in nervousness, fear and cautiousness were labeled as neurotic; the corresponding term in Jones’s study was fearfulness. Those who were persevering, independent, tenacious and assertive were labeled as self-assured/ motivated. Those who were described as obedient, reliable, intelligent and attentive were considered trainable, a personality trait that the Jones study also included. Finally, dogs who were friendly, non-aggressive, sociable and happy-golucky were deemed amicable.

The difference between the Australian study and Jones’s was that the Australian researchers used descriptive adjectives and Jones used behavioral descriptions to concretely define the behaviors of interest, such as aggression. Further steps required in the Australian study include determining whether different observers consistently give the same dog similar ratings, and whether the results can predict how the dog will behave in certain contexts.

The next step in Jones’s research was to determine whether these questionnaire findings would have predictive value. To do this, the same dogs were assessed behaviorally in a doggie day care situation. Dogs were taken into an unfamiliar room and a battery of tests was performed to see how they reacted. For instance, 18 tests measured levels of fear, including the response to someone reaching for the dog’s collar and the approach of an unfamiliar man or woman. Similarly, tests for trainability involved learning new behaviors—researchers gathered information on what each dog found rewarding and then chose a task with which the dog didn’t have much experience. They’d pair a command with, for example, having the dog touch a block and receive a motivator as a reward. They rated how quickly the dog acquired the behavior, how attentive he was during training, how easy it was to motivate him and how distracted he was by the environment. They also looked at commands they would expect most dogs to know, partly to see how attentive the dog was. Overall, Jones found that using behavioral measures validated their results.

These findings show not only that human personality research methods can be adapted for use with dogs, but also, that the canine personality test has promise for use with both shelter dogs and our own companion pups. One of the projects’ goals is to help shelters facilitate better matches with potential adopters, and studies such as these add to the tools they have at their disposal.

On a broader scale, the owner who knows an individual dog’s personality could better shape the relationship they want. For instance, if a dog is inattentive to commands, slow to come when called, quick to sneak out doors and slow to respond to corrections, his personality would be described as low in trainability. Owners of such a dog would do best if they realized that he wasn’t going to be as easy to train as a dog who scored high in that personality trait.

Ideally, one day, humans could be matched with companion animals based on personality tests. Such studies have been performed in some species; for example, Dr. Becky Fox, a researcher at the University of Nevada, Reno, studied the effect of personality on pair-compatibility in cockatiels. She found that agreeableness was the most important factor in determining whether birds would pair. Says Fox, “The more closely the male and female matched on agreeableness, the more compatible they were.” By expanding our knowledge of canine personality, we could have a better idea about which dogs would get along best together in a household, as well as which one would be most in sync with our particular personality. It would add an entirely new dimension to our understanding of dogs.

Article first appeared in The Bark, Issue 51: Nov/Dec 2008

Sophia Yin, DVM, (deceased) was an applied animal behaviorist. A long-time The Bark contributing editor, she was also the author of two behavior books.