Behavioral assessment of child-directed canine aggression
Ilana R Reisner, Frances S Shofer, Michael L Nance
Injury Prevention 2007;13:348–351. doi: 10.1136/ip.2007.015396
Objective: To characterize behavioral circumstances of bites to
children by dogs presented to a veterinary behavior clinic.
Methods: Retrospective case series examining medical records
of dogs presenting by referral to a university veterinary hospital
for aggression and which had bitten a child ,18 years old.
Behavioral data included age of victim, familiarity with dog,
and circumstances of bites.
Results: Records of bites to 111 children were examined.
Children ,6 years old were most commonly bitten in
association with resource guarding (44%), whereas older
children were most commonly bitten in association with territory
guarding (23%). Similarly, food guarding was the most
common circumstance for bites to familiar children (42%) and
territory guarding for bites to unfamiliar children (53%).
Behavioral screening of the 103 dogs examined revealed
resource guarding (61%) and discipline measures (59%) as the
most common stimuli for aggression. Anxiety screens revealed
abnormalities in 77% of dogs. Potential contributory medical
conditions were identified/suspected in 50% of dogs. When
history before presentation was known, 66% of dogs had never
previously bitten a child, and 19% had never bitten any human.
Most dogs (93%) were neutered, and 66% of owners had taken
their dogs to obedience training classes.
Conclusions: Most children were bitten by dogs with no history
of biting children. There is a high rate of behavioral
abnormalities (aggression and anxiety) in this canine popula-
tion. Common calming measures (neutering, training) were not
routinely effective deterrents.
ost dog bites reported to public health authorities are
inflicted on children.
Whereas there are a number of
studies reporting the epidemiologic characteristics of
dog bite injury, information about the behavior of the dog or
bite victim is limited.
Veterinary behavioral medicine is a recently recognized
specialty in veterinary medicine.
Data obtained in a veterinary
behavior clinic can help pediatricians, parents, and other care
givers to better understand the behavioral aspects of child-
directed canine aggression, which, in turn, should lead to more
effective prevention measures.
The records of dogs presenting to the Behavior Clinic of the
Matthew J Ryan Veterinary Hospital of the University of
Pennsylvania (MJR-VHUP) for human-directed aggression
from January 2002 to December 2005 were reviewed retro-
spectively. All cases in which the dog had bitten a child under
the age of 18 years were included. However, bitten children for
whom age or the circumstances of the bite were unknown were
A questionnaire, which included information about the dog,
the owners’ family, description of aggressive incidents as well
as screening questions for aggressive and anxious behaviors
exhibited by the dog, was completed by each dog owner at the
time of the initial appointment. In addition to completed
questionnaires and aggression and anxiety screens, each
medical record included referring veterinary examination and
laboratory results as well as physical examination and
laboratory findings conducted at the time of the veterinary
Bite occurrences were categorized by familiarity of the victim
with the dog and by circumstances surrounding the event.
Familiar children included members of the family and/or
household, or frequent visitors who were not household
members. Unfamiliar children did not live in the household
and were either unknown to the dog or were infrequent visitors
to the home. Circumstances of bites to familiar children
included resource guarding or food guarding, benign (non-
aversive) interaction, aversive but non-painful interaction,
aversive and painful interaction, or interaction while the dog
was resting or sleeping. Circumstances of bites to unfamiliar
children were categorized relative to the dog’s perceived
territory (house, yard, and surrounding area), and as either
interactive or non-interactive.
Data are presented using frequencies and percentages. To
compare children in different age groups by biting circum-
stances, the Fisher exact test was used. Where applicable, data
are presented as differences with 95% CI. Statistical significance
was defined as p,0.05. All analyses were performed using SAS
V.9.1 (SAS Institute, Cary, North Carolina, USA).
A total of 145 children under the age of 18 years were bitten. Of
these, 111 met inclusion criteria. Thirty four children (31%)
were younger than 6 years old, and 77 (69%) were 6–17 years
old. Half were boys and half were girls.
Familiar children were most commonly bitten in relation to
food or resource guarding (n = 29; 26%) and ‘‘benign’’
interactions (n = 20; 18%) such as petting, hugging, bending
over, or speaking to the dog. Presence in or entering the dog’s
territory was the most common situation in which unfamiliar
children were bitten, regardless of whether the child was
(n = 10; 9%) or was not (n = 21; 19%) actively interacting with
the dog. Unfamiliar children were also bitten away from the
dog’s home or yard, regardless of interaction (n = 4; 4%) or lack
of interaction (n = 5; 5%).
Table 1 summarizes the circumstances of the biting episode
relative to both age of the child and familiarity with the dog.
Children ,6 years old were significantly more likely than older
children to be bitten in relation to food guarding or other
resource-associated aggression (44% vs 18%, difference = 26%,
95% CI 4 to 45%, p = 0.009) or in aversive, potentially pain-
eliciting interactions such as stepping or falling on the dog
(18% vs 0%; difference = 18%, 95% CI 7 to 35%, p = 0.0006). No
differences were noted between girls and boys in any of these
A total of 103 dogs had bitten a child under the age of 18 years.
Three quarters of the dogs were male (n = 77; 75%), and all but
four males and three females had been neutered. Forty one
breeds were represented. English Springer Spaniels and
German Shepherd Dogs each comprised 9% of pure-bred dogs
(7% of all dogs), followed by 5% each of Labrador Retrievers,
Golden Retrievers, and American Cocker Spaniels (4% of all
dogs). The total number of times a dog had bitten (historically)
was known for 98 dogs. Nineteen (19%) presented for the
single bite incident involving a child, and had never previously
bitten a person of any age. The remaining 79 (81%) dogs had
bitten at least one person (the same child, or a different child or
adult) more than once (two bites, 15%; three bites, 18%; four
bites, 13%; five bites, 9%; more than five bites, 24%). When the
historical details of bites before the current bite were known,
66% (n = 48) of dogs had never previously bitten a child. Forty
four dogs (45%) had bitten a child who was a member of the
family or household, and 65% had bitten either child or adult
members of the family or household. Thirty five (35%) dogs had
bitten only unfamiliar children. In some cases, histories were
largely unknown because of age at acquisition.
Aggression screens completed by the owner of each dog
revealed that the most common circumstance associated with
aggression historically, to either adults or children, was
resource guarding (61%) (table 2). Similarly, dog anxiety
screening demonstrated common abnormal or reactive beha-
vioral tendencies (table 2).
Table 1 Circumstances of aggression to 111 children by dogs presented to a veterinary behavior service
Category Stimulus description
(n = 34)
(n = 77) p Value
Aggression to familiar children
Resource guarding Approaching, reaching for or touching dog while dog is near or
eating/chewing food, bone or toy
15 (44) 14 (18) 0.009
Benign (non-aversive) Petting, hugging, kissing, bending over, reaching, speaking 5 (15) 15 (19) NS
Resting Waking dog; interacting while dog is resting; approaching while
dog is resting/lying down; lying beside recumbent dog;
pushing/pulling dog off furniture
1 (3) 8 (10) NS
Aversive, painful Stepping on dog; pulling on hair or body; falling on to dog;
trimming nails; punishment by hitting or use of leash correction
6 (18) 0 (0) 0.0006
Aversive, non-painful Restraint/pulling by collar; grooming, toweling feet; bathing;
lifting; verbally scolding
0 (0) 5 (6) NS
Aggression to unfamiliar children
Territorial, non-interactive In/on dog’s territory, including house, yard, area surrounding home,
car; child does not interact
3 (9)* 18 (23) NS
Territorial, interactive In/on dog’s territory, including house, yard, area surrounding home,
car; child interacts, eg, by speaking to dog, petting, reaching,
attempting to feed, removing objects, restraining
2 (6)* 8 (10) NS
Not territorial, interactive Not on dog’s territory; child interacts, eg, by speaking to dog, petting,
reaching, attempting to feed, removing objects, restraining
0 (0) 4 (5) NS
Not territorial, non-interactive Not on dog’s territory; child does not interact 1 (3) 4 (5) NS
Aggression - other 2 (6) 7 (9)
Values are number (%). Columns may add up to more than total because children could have been bitten in more than one context (one child in ,6-year-old group; six
children in 6–17-year-old group).
*Total N = 33; cirumstances of territorial bite were unknown for one child.
Comparison was not performed because bite circumstances varied.
Table 2 Responses to canine aggression and anxiety screens by owners of 103 dogs
presented to a veterinary behavior clinic with a history of biting children
Remove dog food, special food, toys (resource guarding) 48 (61) 31 (39) 24
Punish (verbally scold, correct with leash, hit) 24 (59) 17 (41) 62
Disturb while sleeping or resting; push or pull off furniture 38 (49) 40 (51)25
Reach over or toward dog 34 (38) 56 (62) 13
Bathe, groom, or towel 21 (26) 59 (74) 23
Anxiety related to separation from owner(s) while owner is
34 (35) 62 (65)
Anxiety related to separation from owner(s) while owner is
present but inaccessible
49 (51) 47 (49)
Anxiety or fear related to thunderstorms/fireworks 47 (50) 47 (50)
Any anxiety (either separation or storm/noise-related
anxiety, or both)
78 (77) 23 (23)
Values are number (%) or number. Aggression was directed to children, adults, or both. ‘‘Situation does not apply’’
refers to situations or provocations that do not occur for the particular dog. For example, resource guarding cannot be
assessed if owners have not attempted to remove food or toys. In the aggression screen, positive responses refer to
growling, baring teeth, lunging, snapping, or biting a person in response to the listed stimulus. In the anxiety screen,
positive responses refer to trembling, panting, pacing, vocalization, destructiveness, urination, or defecation. Anxiety
screen rows may not add up to 103 because of missing values.
Behavioral assessment of child-directed canine aggression 349
On the basis of clinical assessment, fear-related aggression
was the most common primary behavioral diagnosis in the dogs
(n = 90; 87%), followed by resource guarding (n = 53; 51%),
territorial defense (n = 52; 51%), social conflict (n = 40; 39%),
and pain (n = 14; 14%). Additional diagnoses included general-
ized anxiety (n = 64; 62%), inappropriate or excessive atten-
tion-seeking behavior (n = 36; 35%), and clinically significant
noise or thunderstorm fear (n = 30; 29%) and separation
anxiety (n = 18; 17%).
On the basis of physical examination, laboratory findings,
and observation, a medical problem was identified or suspected
in 51 (50%) dogs. Orthopedic (n = 18; 20% of all dogs
examined) and dermatologic (n = 18; 20%) conditions were
most commonly identified. Other medical problems included
dermal or epidermal masses and ophthalmologic, metabolic
(eg, renal and hepatic), endocrinologic, and infectious (eg,
Borrelia burgdorferi) disease.
Most owners (66%) had taken their dogs through formal
obedience training classes. Twenty one families had no prior
experience, as adults, with dog ownership; however, prior
experience or its lack had no significant association with biting.
In this study, we describe the circumstances surrounding bites
to children by dogs evaluated for aggressive behavior at a
university-based veterinary behavior service. Although the
epidemiology of bitten children has been reported in a number
of studies, there have been few studies on the circumstances of
aggression, or behavioral or medical information about the
biting dogs themselves. This is the first study to examine the
behavioral aspects of child-directed canine aggression from the
point of view of a veterinary behavioral assessment.
Although 66% of the evaluated dogs had no prior history of
biting children, behavioral abnormalities were universally
present in this canine population. Historically, although 19%
of dogs had never bitten before presentation, a history of
aggressive behavior other than biting (eg, baring teeth) was
common. Furthermore, although some types of human-
directed aggression tend to be observed only in behaviorally
mature dogs (starting at 1–3 years of age),
to food or pain may be seen in juvenile dogs.
even in a puppy, and even in the absence of biting, should raise
concern and consideration should be given to referral for
behavioral evaluation of the dog.
Anxiety screening identified abnormalities in 77% of animals.
Historical evidence of fearful or anxious behavior in response to
loud noises and thunderstorms or separation from the owner
may signal a predisposition to biting in threatening situations
related to anxiety or fear.
Dogs that react with anxiety to
threatening stimuli may be more likely to react aggressively to
children, who, particularly when very young, are at risk because
of their high-pitched voices, sudden movements, and inap-
Medical conditions were identified or suspected in 50% of the
dogs evaluated. There were a number of dogs with orthopedic,
dermatologic, and other disease both at the time of consultation
and historically. These associated medical or painful conditions
may have increased the risk of aggression. However, because of
the lack of a well-animal clinic for comparison of presented
dogs, it was not possible to determine whether this was higher,
lower, or as expected in the patient population. Because disease
and pain can increase irritability in dogs,
even if a causal
relationship is not confirmed, their presence should be an
indication to separate the dog from young children until the
disease has been treated or the pain reduced.
Previous reports of dog bites to children have made safety
recommendations, such as neutering male dogs,
and avoiding specific breeds.
prevalence of males (75%) in our study is similar to other
Almost all dogs, both male and female, were neutered.
Although our data did not include age at neutering or whether
the surgery occurred before or after the appearance of
aggressive behavior, it is apparent that neutering does not
guarantee a reduction of aggression in dogs. It is interesting to
note that the predominant canine behavioral diagnosis, fear-
related aggression, lacks sexual dimorphism,
neither sex should be over-represented. However, even male-
associated aggression such as territorial defense is unlikely to
be eliminated by neutering.
Regardless of neuter status,
parents seeking a pet dog might be advised to seek a female.
Two-thirds of the dogs in this study had been taken to training
classes by their owners. It is not known whether owners had
made specific efforts to train or socialize dogs to be comfortable
with children. Although the success of obedience training for
individual dogs was not measured, the results of this study
suggest that obedience training, like neutering, will not ensure
prevention of future bites to children. However, the efficacy of
obedience training in reducing aggression was not specifically
measured. Cohort studies would be needed to evaluate whether
training (or neutering) reduces biting behavior. With the
exception of the English Springer Spaniel, the breeds included
in our study ranked high in American Kennel Club breed
registrations and appear to reflect breed popularity. Because the
total number of English Springer Spaniels in our study was
small, and the study was performed at a referral hospital with a
highly selected group of patients, it is safest to conclude that
any breed of dog is capable of biting a child.
The findings for younger children were not unexpected. Food
or resource guarding is a common behavior problem in dogs
and was noted in almost two thirds of the dogs in this study.
To be safe, children of any age should not be permitted near the
dog whenever food (including human food) is present.
The meaning of ‘‘provocation’’ has been inconsistent in the
literature and should be interpreted with caution.
presence of a parent, who may underestimate the risk of bites
to young children,
may not be sufficient to prevent bites.
Although it is natural to assume that hitting and other pain-
inducing interactions can elicit aggression, parents and dog
owners may be less vigilant when a child simply approaches or
pets a dog. Similarly, for unfamiliar children, walking or cycling
near a dog’s home may be provocative enough when dogs are
tethered outdoors or are not securely fenced.
Our study focuses on children bitten by pet dogs evaluated in
a secondary and tertiary care veterinary behavior clinic with a
history of aggression to children. We recognize the limitations
of a retrospective case series study at a referral center. Our
patients are a highly selected group of dogs, and the ability to
Children are at risk of dog bite in association with
resource guarding by the dog or pain-causing interac-
tions. Unfamiliar children are at risk of dog bite while in
the dog’s home, yard, or perceived territory, regardless
of whether or not they are interacting with the dog.
All dogs evaluated for human-directed aggression in our
study had a behavior and/or medical abnormality.
Episodes of aggression were not limited to specific dog
breeds, gender, neuter status, or history of training.
The risk of biting may be increased in the presence of
pain or disease in the pet.
350 Reisner, Shofer, Nance
draw generalizable inferences from them is limited. We also
acknowledge the limitations of a retrospective study of self-
report and self-assessment (of their pets) by dog owners
seeking help for problem behavior. However, we do attempt to
better characterize this common clinical problem from the
unique perspective of the canine behavioral analysis.
We thank Ms Alison Seward and Ms Jenny O’Connor, CVT, for their
assistance in data collection.
Ilana R Reisner, Frances S Shofer, Department of Clinical Studies, School
of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA, USA
Michael L Nance, Department of Surgery, The Children’s Hospital of
Philadelphia, PA, USA
Competing interests: None.
Correspondence to: Dr I R Reisner, Department of Clinical Studies, School
of Veterinary Medicine, University of Pennsylvania, 3900 Delancey St,
Philadelphia, PA 19104-6010, USA; email@example.com
Accepted 26 June 2007
1 Mathews JR, Lattal KA. A behavioral analysis of dog bites to children. J Dev
Behav Pediatr 1994;15:44–52.
2 Gershman KA, Sacks JJ, Wright JC. Which dogs bite? A case-control study of
risk factors. Pediatrics 1994;93:913–17.
3 Kahn A, Bauche P, Lamoureux J. Child victims of dog bites treated in emergency
departments: a prospective survey. Eur J Pediatr 2003;162:254–8.
4 Avner JR, Baker MD. Dog bites in urban children. Pediatrics 1991;88:55–7.
5 European College of Veterinary Behavioural Medicine. Companion animal.
http://www.ecvbm.org/ (accessed 20 Jul 2007).
6 American College of Veterinary Behaviorists. http://www.DACVB.org/
(accessed 20 Jul 2007).
7 Reisner IR. Differential diagnosis and management of human-directed
aggression in dogs. Vet Clin North Am Small Anim Pract 2003;33:303–20.
8 Borchelt PL, Voith VL. Aggressive behavior in dogs and cats. Compend Contin
Educ Pract Vet 1985;7:949–57.
9 Overall KL. Medical differentials with potential behavioral manifestations. Vet
Clin North Am Small Anim Pract 2003;33:213–29.
10 Mertens PA. Canine aggression. In: Horwitz D, Mills D, Heath S, eds. BSAVA
Manual of canine and feline behavioural medicine. Gloucester: British Small
Animal Veterinary Association, 2002:195–215.
11 Pryor P. Animal behavior case of the month: an English Cocker Spaniel was
examined because of growling and snapping at people and dogs. J Am Vet Med
12 Bandow JH. Will breed-specific legislation reduce dog bites? Can Vet J
13 Presutti RJ. Prevention and treatment of dog bites. Am Fam Physician
14 Reisner I. National survey of owner-directed aggression in English Springer
Spaniels. J Am Vet Med Assoc 2005;227:1594–603.
15 Guy NC, Luescher UA, Dohoo SE, et al. A case series of biting dogs:
characteristics of the dogs, their behaviour, and their victims. Appl Anim Behav
16 Gandhi RR, Liebman MA, Stafford BL, et al. Dog bite injuries in children: a
preliminary survey. Am Surg 1999;65:863–4.
17 Patrick GR, O’Rourke KM. Dog and cat bites: epidemiologic analyses suggest
different treatment strategies. Public Health Rep 1998;113:252–7.
18 Bernardo LM, Gardner MJ, Rosenfield RL, et al. A comparison of dog bite injuries
in younger and older children treated in a pediatric emergency department.
Pediatr Emerg Care 2002;18:247–9.
19 Lauer EA, White WC, Lauer BA. Dog bites: a neglected problem in accident
prevention. Am J Dis Child 1982;136:202–4.
20 Villar RG, Connick M, Barton LL, et al. Parent and pediatrician knowledge,
attitudes, and practices regarding pet-associated hazards. Arch Pediatr Adolesc
21 Chun Y-T. Dog bites in children less than 4 years old. Pediatrics
22 Sacks JJ, Sattin RW, Bonzo SE. Dog bite-related fatalities from 1979 through
1988. JAMA 1989;262:1489–92.
Webcast: International Forum on Quality and Safety in Health Care
Plenary sessions at this year’s International Forum on Quality and Safety in Health Care were
filmed and broadcast live over the internet. The sessions are still available to view free, on demand
and at your own convenience at http://barcelona.bmj.com. Each session is accompanied by a
The webcast includes the following, in either English or Spanish translation:
Donald M Berwick: Can health care ever be safe?
Richard Smith: What the quality movement can learn from other social movements
Lucian Leape and Linda Kenney: When things go wrong: communicating about adverse events
John Prooi and Harry Molendijk: Partnering for patient safety
Behavioral assessment of child-directed canine aggression 351