Publishing Authors : Deborah L. Duffy , Yuying Hsu b, James A. Serpell
Date Published : 2008
Canine aggression poses serious public health and animal welfare concerns. Most of what is understood
about breed differences in aggression comes from reports based on bite statistics, behavior clinic caseloads,
and experts’ opinions. Information on breed-specific aggressiveness derived from such sources may be
misleading due to biases attributable to a disproportionate risk of injury associated with larger and/or more
physically powerful breeds and the existence of breed stereotypes. The present study surveyed the owners of
more than 30 breeds of dogs using the Canine Behavioral Assessment and Research Questionnaire (CBARQ),
a validated and reliable instrument for assessing dogs’ typical and recent responses to a variety of
common stimuli and situations. Two independent data samples (a random sample of breed club members
and an online sample) yielded significant differences among breeds in aggression directed toward strangers,
owners and dogs (Kruskal–Wallis tests, P < 0.0001).
Eight breeds common to both datasets (Dachshund, English Springer Spaniel, Golden Retriever,
Labrador Retriever, Poodle, Rottweiler, Shetland Sheepdog and Siberian Husky) ranked similarly,
rs = 0.723, P < 0.05; rs = 0.929, P < 0.001; rs = 0.592, P = 0.123, for aggression directed toward strangers,
dogs and owners, respectively. Some breeds scored higher than average for aggression directed toward both
humans and dogs (e.g., Chihuahuas and Dachshunds) while other breeds scored high only for specific targets
(e.g., dog-directed aggression among Akitas and Pit Bull Terriers). In general, aggression was most severe
when directed toward other dogs followed by unfamiliar people and household members. Breeds with the
greatest percentage of dogs exhibiting serious aggression (bites or bite attempts) toward humans included
Dachshunds, Chihuahuas and Jack Russell Terriers (toward strangers and owners); Australian Cattle Dogs
(toward strangers); and American Cocker Spaniels and Beagles (toward owners). More than 20% of Akitas,
Jack Russell Terriers and Pit Bull Terriers were reported as displaying serious aggression toward unfamiliar
dogs. Golden Retrievers, Labradors Retrievers, Bernese Mountain Dogs, Brittany Spaniels, Greyhounds and
Whippets were the least aggressive toward both humans and dogs. Among English Springer Spaniels,conformation-bred dogs were more aggressive to humans and dogs than field-bred dogs (stranger
aggression: Mann–Whitney U test, z = 3.880, P < 0.0001; owner aggression: z = 2.110, P < 0.05; dogdirected
aggression: z = 1.93, P = 0.054), suggesting a genetic influence on the behavior. The opposite
pattern was observed for owner-directed aggression among Labrador Retrievers, (z = 2.18, P < 0.05)
indicating that higher levels of aggression are not attributable to breeding for show per se.
Canine aggression presents serious public health, economic and animal welfare concerns.
Recent reports estimate that hospital emergency rooms treat over 300,000 dog bite injuries per
year in the USA, nearly half of which involve children under the age of 15 years. Approximately
2–4% of all dog bite cases require hospitalization (Weiss et al., 1998; Centers for Disease Control
and Prevention, 2003).
In a study of 12 different animal shelters across the USA, 40% of relinquishing owners cited
behavioral problems as one of the reasons for surrendering a dog. When behavior was the only
reported reason for relinquishment, aggression was the most frequently cited problem (40% of
dogs) (Salman et al., 1998).
Four primary approaches have been used to investigate breed-related variation in aggressive
behavior: analyses of dog bite statistics (e.g., Gershman et al., 1994; Lockwood, 1995; Guy et al.,
2001b; Reisner et al., 2005), behavior clinic/consultant caseloads (e.g., Beaver, 1983; Borchelt,
1983; Blackshaw, 1991; Bamberger and Houpt, 2006), the opinions of dog experts (veterinarians
and trainers; e.g., Hart and Hart, 1985; Bradshaw and Goodwin, 1998; Takeuchi and Mori, 2006;
Notari and Goodwin, 2007) and the results of behavioral testing (e.g., Svartberg, 2006). Each
method has its drawbacks.
Dog bite statistics are potentially misleading for several reasons: (a) most dog bites go
unreported unless medical attention is sought (which may be more likely with larger breeds that
have the ability to inflict more serious injury); (b) the total number of dogs of a given breed in the
local community is seldom known, so the degree to which that breed is over-represented among
reported dog bites is usually undetermined (Lockwood, 1995; however see Gershman et al.,
1994; Guy et al., 2001b; Reisner et al., 2005); and (c) in many cases the breed of dog involved
cannot be verified (Wright, 1991).
Breed-specific data on aggression derived from behavior clinic/consultant caseloads are also
likely to be unrepresentative. Because of the greater risk of injury posed by larger, more powerful
dogs, owners of these dogs are more likely to seek professional help in dealing with canine
aggression. In addition, dog owners dealing with aggression directed toward themselves or
members of their family are more likely to seek professional help compared to pet owners whose
dogs are aggressive toward unfamiliar people or animals (Bamberger and Houpt, 2006).
Hart and Hart (1985) pioneered the use of animal ‘experts’ (e.g., veterinarians and obedience
judges) as sources of information on the prevalence of various desirable and undesirable
behaviors among popular dog breeds. Their methods, which involve asking experts to rank
breeds on a series of traits, have also been applied to studies of breed differences in countries
other than the USA, e.g., the UK (Bradshaw and Goodwin, 1998), Italy (Notari and Goodwin,
2007), and Japan (Takeuchi and Mori, 2006), and have provided evidence of agreement regarding
442 D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460
levels of aggression in some breeds. Unfortunately, with surveys of this type, it is difficult to
ascertain whether agreement among experts about the aggressiveness of a particular breed, either
within or between countries, reflects true consistencies in breed behavior or simply shared
In contrast to opinions based upon breed generalizations, assessments directed at individual
dogs may be less susceptible to bias related to breed stereotypes (Notari and Goodwin, 2007).
Such assessments include behavioral tests and dog owner surveys. Behavioral testing for
aggression involves exposing dogs to one or more
For the purposes of analysis, subscale scores are calculated as the average of the scores for the
questionnaire items comprising each factor. In the present paper, we focus on four subscales related to
aggression. The aggression subscales address aggressive responses of dogs to different targets: unfamiliar
people (stranger-directed aggression, SDA), familiar people (owner-directed aggression, ODA), unfamiliar
dogs (dog-directed aggression, DDA), and familiar dogs living in the same household (dog rivalry,
DR). Because fearful responses are often associated with the expression of aggressive behavior in dogs
(Lindsay, 2001, pp. 171–172), we included two additional subscales that address fearful responses directed
toward unfamiliar people (stranger-directed fear) and unfamiliar dogs (dog-directed fear) for some
Calculation of Cronbach’s alpha coefficients (Cronbach, 1951) for the individual aggression subscales
obtained coefficients greater than 0.80, indicating a high level of internal reliability for these scales
(Nunnally, 1978, p. 245; Carmines and Zeller, 1979, p. 51). Similarly, independent (i.e. without conferring)
C-BARQ ratings of a sample of 75 dogs by pairs of owners (e.g., spouses or partners) have also confirmed
that the aggression subscales have good inter-rater reliability characteristics. Average weighted Kappa
coefficients ranged from 0.4 to 0.6 (P < 0.0001), which is considered to be a moderate to substantial degree
of inter-rater agreement (Landis and Koch, 1977). The average weighted percent agreement between pairs of
raters exceeded 90% for all aggression subscales. Intraclass correlation coefficients (ICC) (Bartko, 1966;
Shrout and Fleiss, 1979) of the calculated subscales also indicate a moderate to high degree of inter-rater
reliability (SDA ICC = 0.84; ODA ICC = 0.91; DDA ICC = 0.69; DR ICC = 0.60). Using a sample of 200
clinical referral cases, the construct validity of these subscales was established by confirming their
effectiveness at discriminating between dogs independently diagnosed as either displaying or not displaying
corresponding behavior problems (Hsu and Serpell, 2003).
2.2.1. Breed club sample
Study participants consisted of members of 11 American Kennel Club (AKC) recognized national breed
clubs (see Serpell and Hsu, 2005 for details). C-BARQ questionnaires were distributed to breed club
members by ordinary mail together with an explanatory letter, and a pre-paid return envelope. In an effort to
‘randomize’ the samples, recipients from most of the clubs were selected from either the first or last 300
members listed alphabetically in each club’s membership directory. The Labrador Retriever Club elected to
distribute the C-BARQ to its own members, and sent questionnaires to the first 488 members listed
alphabetically in its directory. The English Springer Spaniel Field Trial Association also chose to distribute
the C-BARQ: all 187 members with field (working)-bred Spaniels were sent questionnaires, and the Trinity
College (Dublin, Ireland) online random number generator (http://random.org/) was used as the basis for
sampling 300 out of a total of 367 members with conformation (show)-bred Spaniels. In addition to the CBARQ
assessments, information was also collected on each dog’s age, sex and neuter status. For the
Labrador and English Springer Spaniel breeds, owners were also asked to state whether the dog was field or
conformation (show)-bred, if known.
To ensure statistical independence, each respondent was asked to assess only one dog, preferably one that
was well known, that was at least 1 year old at the time of assessment in order to reduce maturational effects
(Serpell and Jagoe, 1995).
2.2.2. Online sample
Beginning in April 2006, free access to an online version of the C-BARQ (http://www.vet.upenn.edu/
cbarq) became available to pet owners. The online survey was advertised via an article in the newsmagazine
of the Veterinary Hospital of the University of Pennsylvania, USA (http://www.vet.upenn.edu/bellwether/
v64/article10.shtml) and by notices sent to Philadelphia-area veterinary clinics and the top 20 USA breed
clubs based on AKC registrations. Availability of the survey then spread via word of mouth. This sample of
dog owners is therefore self-selected which we note as a potential source of bias. Breed designations are
based entirely upon owner assertions.
444 D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460
Owners were permitted to complete questionnaires for as many dogs as they wished. However, for data
analysis we used a random number generator (available as part of the statistics software, SPSS 15.0, SPSS,
Inc.) to randomly select only one dog per owner to ensure statistical independence. Only breeds for which at
least 45 dogs were surveyed were included in analyses.
Data were analysed using SPSS 15.0 for Windows (SPSS, Inc.). Due to non-normality of the data that
could not be corrected with transformation, non-parametric tests were used to analyse differences in CBARQ
subscale scores. Mann–Whitney U and Kruskal–Wallis tests were used for between-group
comparisons of continuous variables (age and aggression subscale scores). Chi-square tests of independence
were used for analyses of sex ratios and frequency of neutered/spayed dogs. To distinguish dogs that tended
to engage in aggressive displays (e.g., barking, growling, etc.) from those that actually bit or attempted to
bite, the proportion of dogs in each breed that scored at least one ‘‘4’’ (‘‘snaps, bites or attempts to bite’’) on
items pertaining to each of the aggression subscales was calculated. Because inter-dog aggression within a
household can only occur if there is more than one dog, ‘dog rivalry’ cases that had missing values for all
four questions that comprise this subscale were excluded. Spearman rank order correlation coefficients were
used to assess associations between breed means for C-BARQ subscale scores and percentages reported as
Correlations among the subscales were determined using Spearman rank order correlation coefficients.
Partial correlations were used to assess the relationship between ODA and DR while controlling for DDA
and SDA. To assess breed differences in aggressiveness relative to fearfulness, we subtracted the population
mean from the mean of each breed for stranger- and dog-directed aggression and fear.
To assess how breeds common to both the online and breed club datasets compared to one another,
Spearman rank order correlation coefficients were used to compare mean subscale scores.
3.1. Descriptive statistics
3.1.1. Breed club sample
A total sample of 1553 complete C-BARQs were returned (average return rate 49%). Twentyfour
cases were removed due to missing data regarding the dog’s sex (three cases) or age (21
cases). Despite our instructions, a small number of dogs were less than 1 year old (n = 27). Dogs
that were less than 6 months old (five cases) or greater than 17 years old (three cases) were
excluded from the data to eliminate possible effects of extreme immaturity and senility,
respectively. Table 1 summarizes the primary descriptive characteristics of the sample. Breeds
did not differ significantly from one another with respect to sex ratio but there were significant
breed differences in the ratio of intact to neutered dogs (x2 = 60.81, d.f. = 10, P < 0.0001).
Overall, there were more intact dogs compared to neutered dogs among most breeds. Breeds also
differed from one another in age (Kruskal–Wallis H = 63.83, d.f. = 10, P < 0.0001).
3.1.2. Online sample
A total of 8260 complete C-BARQs were available as of December 3, 2007. Of those, 1257
dogs whose breed status was reported as ‘mixed/unknown’ were removed from the data.
Removing breeds with fewer than 45 dogs (2051 dogs representing 143 breeds) left 4952 dogs.
Random selection of one dog for each owner provided 3791 C-BARQs that were included in
analyses, representing 33 different breeds (Table 2). There were significant breed differences for
sex ratio (x2 = 48.97, d.f. = 32, P < 0.05) and in the ratio of intact to neutered dogs (x2 = 251.84,
Descriptive statistics of the breed club sample of dogs used in the study
Breed n Female (%) Neutered (%) Age in years
Basset Hound 151 52.98 50.99 5.65 3.10
Dachshund 120 47.50 44.17 6.58 3.46
English Springer Spaniel 247 51.01 27.94 5.06 2.42
Golden Retriever 179 46.37 49.72 5.29 2.83
Labrador Retriever 277 53.43 25.63 5.68 2.83
Poodlea 69 60.89 34.78 6.87 3.43
Rottweiler 92 50.00 41.30 6.00 2.94
Shetland Sheepdog 112 47.32 49.11 7.48 3.79
Siberian Husky 92 42.39 42.39 7.31 4.08
West Highland White Terrier 92 60.87 44.57 6.66 3.64
Yorkshire Terrier 90 60.00 44.44 6.12 3.02
All breeds 1521 51.55 39.18 5.99 3.20
a Includes standard, miniature, and toy varieties.
(rs = 0.311, n = 1356, P < 0.0001), although breeds differed in their patterns of aggression
relative to fear directed at strangers and dogs (Fig. 2A and B). Some breeds (e.g., Dachshund)
showed high levels of both behaviors (Fig. 2A and B), some tended to be more aggressive than
fearful, particularly in relation to unfamiliar dogs (e.g., West Highland White Terriers and
Rottweilers, Fig. 2B), while none were markedly more fearful than aggressive.
3.2.2. Online sample
Significant differences were observed across the 33 breeds for subscale scores for each of the
aggression subscales (Kruskal–Wallis H = 383.21, d.f. = 32, P < 0.0001; H = 132.76, d.f. = 32,
P < 0.0001; H = 306.93, d.f. = 32, P < 0.0001; H = 173.52, d.f. = 32, P < 0.0001 for SDA,
D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460 447
Descriptive statistics of the online sample of dogs used in the study
Breed n Female (%) Neutered (%) Age in years
Airedale Terrier 66 51.5 75.8 4.39 2.88
Akita 99 42.4 63.6 4.44 2.98
Australian Cattle Dog 136 47.1 86.8 4.22 3.35
Australian Shepherd 177 50.3 76.8 4.23 2.95
Beagle 63 44.4 85.7 5.06 3.65
Bernese Mountain Dog 67 46.3 76.1 3.87 2.37
Bichon Frise 65 50.8 90.8 4.32 3.14
Border Collie 163 49.7 79.1 4.41 3.06
Boxer 70 47.1 82.9 4.13 2.87
Brittany Spaniel 66 42.4 86.4 4.54 3.34
Chihuahua 56 50.0 82.1 4.27 3.28
Cocker Spaniel (American) 107 44.9 81.3 4.23 3.30
Collie 132 48.5 51.5 4.72 3.19
Dachshund 68 41.2 80.9 4.33 3.55
Doberman Pinscher 144 59.0 56.9 4.51 2.97
English Springer Spaniel 57 31.6 73.7 4.57 3.64
German Shepherd 292 53.1 71.9 4.03 3.10
Golden Retriever 181 48.6 81.2 4.56 3.18
Great Dane 53 54.7 75.5 3.87 2.92
Greyhound 62 54.8 100 5.48 3.13
Havanese 73 49.3 63.0 3.05 2.40
Jack Russell Terrier 78 39.7 76.9 5.33 3.72
Labrador Retriever 349 50.4 79.9 4.31 3.16
Mastiff (English) 126 31.7 52.4 2.27 1.70
Pit Bulla 132 49.2 88.6 3.98 3.20
Poodleb 169 42.6 75.7 4.71 3.42
Portuguese Water Dog 75 61.3 68.0 4.41 2.76
Rhodesian Ridgeback 69 46.4 52.2 4.56 3.67
Rottweiler 210 47.1 55.2 3.74 2.56
Shetland Sheepdog 57 50.9 82.5 5.42 3.19
Siberian Husky 54 40.7 72.2 4.40 3.33
Soft Coated Wheaten Terrier 216 48.1 78.2 5.12 3.30
Whippet 59 49.2 71.2 4.93 3.54
All breeds 3791 47.4 75.0 4.33 3.15
a Includes American Pit Bull Terriers, American Staffordshire Terriers and Staffordshire Bull Terriers. b Includes standard, miniature, and toy varieties.
ODA, DDA and DR, respectively; Fig. 3A–D). Large effect sizes between subscale scores for the
least and most aggressive breeds were observed for SDA (Cohen’s d = 1.80; Dachshund vs.
Siberian Husky, Fig. 3A), DDA (d = 1.16; Akita vs. Greyhound, Fig. 3C) and DR (d = 0.98;
Chihuahua vs. Brittany Spaniel, Fig. 3D); a medium effect size was found for ODA (d = 0.60;
Dachshund vs. Rhodesian Ridgeback, Fig. 3B). The breed-specific prevalence of bites or bite
attempts was examined as previously described (Table 4). As with the breed club sample, breed
average C-BARQ aggression scores were significantly positively correlated with the tendency to
bite (SDA: rs = 0.709, n = 33, P < 0.0001; ODA: rs = 0.482, n = 33, P < 0.005; DDA:
rs = 0.921, n = 33, P < 0.0001; DR: rs = 0.685, n = 33, P < 0.0001).
Significant correlations were found between subscale scores for aggression and fear directed
toward strangers (rs = 0.409, n = 3216, P < 0.0001) and dogs (rs = 0.316, n = 3003, P < 0.0001),
and breeds differed in their patterns of aggression relative to fear directed at strangers and dogs
(Fig. 4A and B). Again, some breeds (e.g., Dachshund and Chihuahua) displayed exceptionally
high levels of aggression and fear (Fig. 4A and B), some were more aggressive than fearful,
particularly with respect to other dogs (e.g., Akita, Jack Russell Terrier and Pit Pull, Fig. 4B) and
only a minority was more fearful than aggressive (e.g., Greyhound and Shetland Sheepdog,
448 D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460
Fig. 1. Mean scores (95% confidence intervals) for (A) stranger-, (B) owner- and (C) dog-directed aggression for each
of the 11 dog breeds from the breed club survey. Horizontal bars indicate the population means.
Fig. 4A and B). A significant correlation was also found between subscale scores for ODA and
DR (rs = 0.385, n = 2448, P < 0.0001) that remained significant when controlling for SDA and
DDA (rs = 0.326, n = 2446, P < 0.0001).
3.2.3. Comparisons between samples
Limiting the analysis to the eight breeds common to both samples, small but significant
differences were found between subscale scores for the breed club sample (n = 1186) and the
online sample (n = 1045) for SDA (Cohen’s d = 0.05; Mann–Whitney U test, z = 4.86,
P < 0.0001), ODA (d = 0.03; z = 2.85, P < 0.025) and DDA (d = 0.14; z = 2.35, P < 0.025),
with dogs from the breed club sample scoring somewhat higher for SDA and lower for the other
D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460 449
Number of dogs of each breed from the breed club survey that received the maximum score of 4 for aggression (snaps,
bites, or attempts to bite) for at least one question pertaining to each subscale
‘‘Snaps, bites, or attempts to bite’’
n Stranger aggr. Owner aggr. Dog aggr.
Basset Hound 151 1 (0.66) 4 (2.65) 3 (1.99)
Dachshund 120 8 (6.67) 1 (0.83) 5 (4.17)
English Springer Spaniel 247 3 (1.21) 6 (2.43) 11 (4.45)
Golden Retriever 179 0 (0) 0 (0) 1 (0.56)
Labrador Retriever 277 2 (0.72) 0 (0) 2 (0.72)
Poodle 69 3 (4.35) 0 (0) 0 (0)
Rottweiler 92 5 (5.43) 1 (1.09) 1 (1.09)
Shetland Sheepdog 112 3 (2.68) 1 (0.89) 1 (0.89)
Siberian Husky 92 0 (0) 0 (0) 1 (1.09)
West Highland WhiteTerrier 92 0 (0) 0 (0) 1 (1.09)
Yorkshire Terrier 90 7 (7.78) 1 (1.11) 5 (5.56)
Average 138 3 (2.68) 1 (0.82) 3 (1.96)
Within-breed percentage is given in parentheses.
Fig. 2. (A) Stranger- and (B) dog-directed aggression plotted against stranger- and dog-directed fear for the 11 dog breeds
from the breed club survey. Data points are breed averages relative to the mean scores for all breeds combined.
two forms of aggression compared to dogs from the online sample. Only the difference for SDA
remained significant when the analysis was limited to intact dogs (d = 0.21; z = 4.00,
P < 0.0001; online sample n = 981, breed club sample n = 917).
Significant correlations were observed between the two datasets when the breed averages for
SDA and DDA were compared (rs = 0.723, n = 8, P < 0.05; rs = 0.927, n = 8, P < 0.001,
respectively). The correlation for ODA approached but did not reach significance (rs = 0.592,
n = 8, P = 0.123) (Table 5).
3.3. Conformation vs. field stock
Among English Springer Spaniels in the breed club sample, conformation-bred dogs scored
significantly worse for SDA (Mann–Whitney U test, z = 3.820, P < 0.0001), ODA (z = 2.012,
P < 0.05) and DDA (z = 1.839, P = 0.066) compared with field-bred dogs (Fig. 5A). Labrador
Retrievers showed the opposite pattern for ODA (z = 2.18, P < 0.05) with conformation-bred
dogs scoring lower than field stock. The remaining two aggression subscales revealed no
significant differences among Labrador Retrievers (Fig. 5B).
450 D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460
Fig. 3. Mean scores (95% confidence intervals) for (A) stranger-, (B) owner- and (C) dog-directed aggression and (D)
dog rivalry for each of the 33 breeds of dog from the online survey. Horizontal bars indicate the population means.
4.1. Breed differences in aggression
These findings demonstrate considerable variation among breeds in the prevalence and
severity of aggression directed at different targets (strangers, owners, or other dogs). Although
small differences were observed between the breed club and online samples, breeds were
remarkably consistent relative to one another. To our knowledge, this is the first study to report
replicated findings of breed differences in aggression using the same measure in two independent
samples. Average subscale scores for each breed were significantly correlated with the proportion
showing serious aggression (e.g., biting, snapping), indicating that the C-BARQ subscale scores
provided a reasonably accurate reflection of the relative risks of biting. The findings also suggest
that, for the purpose of obtaining information on the prevalence of behavior problems in the pet
dog population, internet data collection methods provided results comparable to those obtained
by more traditional paper-and-pencil surveys (Gosling et al., 2004).
The present findings should be interpreted with caution. The substantial within-breed
variation in C-BARQ scores observed in this study suggests that it is inappropriate to make
D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460 451
Fig. 3. (Continued).
predictions about a given dog’s propensity for aggressive behavior based solely on its breed.
Furthermore, questionnaire reports inevitably involve a degree of subjectivity, and it is possible
that respondents’ answers were influenced by both popular breed stereotypes and/or perceptions
of which answers would be deemed socially acceptable. The various C-BARQ items are designed
to reduce systematic biases of this kind by focusing on the dog’s recent responses to specific
stimuli and situations. However, in practice, it is impossible to eliminate such biases entirely
using survey methods (Nederhof, 1985). In addition, both the breed club and online samples will
452 D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460
Number of dogs of each breed from the online sample that received the maximum score of 4 for aggression (snaps, bites,
or attempts to bite) for at least one question pertaining to each subscale
‘‘Snaps, bites or attempts to bite’’
n Stranger aggr. Owner aggr. Dog aggr. Dog rivalry N (DR)a
Airedale Terrier 66 3 (4.5) 1 (1.5) 6 (9.1) 2 (3.6) 56
Akita 99 3 (3) 1 (1) 29 (29.3) 5 (7.1) 70
Australian Cattle Dog 136 13 (9.6) 2 (1.5) 28 (20.6) 5 (4.9) 103
Australian Shepherd 177 11 (6.2) 1 (0.6) 26 (14.7) 10 (6.5) 155
Beagle 63 5 (7.9) 5 (7.9) 6 (9.5) 4 (8) 50
Bernese Mountain Dog 67 1 (1.5) 2 (3) 3 (4.5) 1 (1.6) 61
Bichon Frise 65 3 (4.6) 1 (1.5) 3 (4.6) 2 (4.2) 48
Border Collie 163 13 (8) 3 (1.8) 22 (13.5) 7 (5.1) 137
Boxer 70 4 (5.7) 0 (0) 11 (15.7) 3 (5.3) 57
Brittany Spaniel 66 0 (0) 1 (1.5) 3 (4.5) 1 (2) 50
Chihuahua 56 9 (16.1) 3 (5.4) 10 (17.9) 2 (4.8) 42
Cocker Spaniel (American) 107 5 (4.7) 6 (5.6) 8 (7.5) 4 (5.2) 77
Collie 132 2 (1.5) 3 (2.3) 9 (6.8) 2 (1.6) 122
Dachshund 68 14 (20.6) 4 (5.9) 12 (17.6) 5 (8.8) 57
Doberman Pinscher 144 8 (5.6) 2 (1.4) 16 (11.1) 4 (3.4) 119
English Springer Spaniel 57 2 (3.5) 2 (3.5) 10 (17.5) 4 (9.5) 42
German Shepherd 292 13 (4.5) 6 (2.1) 48 (16.4) 1
have been subject to self-selection biases that may have influenced the current findings. On the
other hand, and in spite of these potential limitations, most of our findings were reasonably
consistent with previous reports of breed differences in aggression (Borchelt, 1983; Hart and
Hart, 1985; Wright and Nesselrote, 1987; Bradshaw and Goodwin, 1998; Svartberg, 2006;
Takeuchi and Mori, 2006).
Although some breeds appeared to be aggressive in most contexts (e.g., Dachshunds,
Chihuahuas and Jack Russell Terriers), others were more specific. Aggression in Akitas, Siberian
Huskies, and Pit Bull Terriers, for instance, was primarily directed toward unfamiliar dogs. These
findings suggest that aggression in dogs may be relatively target specific, and that independent
D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460 453
Fig. 4. (A) Stranger- and (B) dog-directed aggression plotted against stranger- and dog-directed fear for the 33 breeds
from the online survey. Data points are breed averages relative to the mean scores for all breeds combined. ((1) Airedale
Terrier, (2) Akita, (3) Australian Cattle Dog, (4) Australian Shepherd, (5) Beagle, (6) Bernese Mountain Dog, (7) Bichon
Frise, (8) Border Collie, (9) Boxer, (10) Brittany Spaniel, (11) Chihuahua, (12) Cocker Spaniel (American), (13) Collie,
(14) Dachshund, (15) Doberman Pinscher, (16) English Springer Spaniel, (17) German Shepherd, (18) Golden Retriever,
(19) Great Dane, (20) Greyhound, (21) Havanese, (22) Jack Russell Terrier, (23) Labrador Retriever, (24) Mastiff
(English), (25) Pit Bull, (26) Poodle, (27) Portuguese Water Dog, (28) Rhodesian Ridgeback, (29) Rottweiler, (30)
Shetland Sheepdog, (31) Siberian Husky, (32) Soft Coated Wheaten Terrier, and (33) Whippet).
Mean scores of aggression for dog breeds common to the breed club and online surveys
Breed Stranger-directed Owner-directed Dog-directed
Online Breed club Online Breed club Online Breed club
Dachshund 1.37 (1) 0.88 (1) 0.27 (1) 0.13 (2) 1.46 (1) 0.89 (1)
English Springer Spaniel 0.60 (2) 0.51 (5) 0.21 (2) 0.19 (1) 1.19 (2) 0.79 (2)
Golden Retriever 0.23 (7) 0.36 (7) 0.09 (6) 0.05 (7) 0.63 (6) 0.49 (6)
Labrador Retriever 0.41 (5) 0.36 (6) 0.07 (8) 0.02 (8) 0.59 (7) 0.46 (7)
Poodle 0.55 (3) 0.65 (4) 0.07 (7) 0.13 (3) 0.81 (3) 0.64 (4)
Rottweiler 0.55 (4) 0.84 (2) 0.11 (4) 0.09 (5) 0.72 (5) 0.72 (3)
Shetland Sheepdog 0.37 (6) 0.67 (3) 0.09 (5) 0.08 (6) 0.51 (8) 0.41 (8)
Siberian Husky 0.07 (8) 0.14 (8) 0.12 (3) 0.09 (4) 0.75 (4) 0.63 (5)
Rankings are in parentheses.
mechanisms may mediate the expression of different forms of aggression. Recent heritability
analyses of aggression in a population of Dutch Golden Retrievers found a weak correlation
between estimated breeding values for C-BARQ ratings of stranger- and dog-directed
aggression, suggesting that these traits are partially related but genetically distinct (Liinamo
et al., 2007).
4.1.1. Stranger-directed aggression (SDA)
There are few published reports describing breed variation in the degree of aggression directed
toward strangers that do not rely on bite statistics. The most extensive and often cited report is a
USA-based survey of veterinarians and obedience judges in which respondents were asked to
rank several popular breeds based on various behavioral traits (Hart and Hart, 1988; see also
Bradshaw and Goodwin, 1998; Takeuchi and Mori, 2006; Notari and Goodwin, 2007). The
respondents’ rankings for each behavior were then transformed into deciles, each containing five
or six breeds, with higher deciles indicating more aggressive behavior. Two behavioral traits,
‘watchdog barking’ (barking to alert owners to an intruder) and ‘territorial defense’ (attacking an
intruder) would be most relevant to our factor of stranger-directed aggression. Several of the
breeds in our study found to be rated highest for stranger-directed aggression (Dachshunds,
Chihuahuas, Doberman Pinschers, Rottweilers, Yorkshire Terriers and Poodles) scored in the
eighth decile or higher for ‘watchdog barking’ and/or ‘territorial defense’ in Hart and Hart (1988)
Breeds scoring low (below the median) for SDA in our study include Basset Hounds,
Golden Retrievers, Labrador Retrievers, Siberian Huskies, Bernese Mountain Dogs, Brittany
Spaniels, Whippets and Greyhounds. Four of these breeds (Basset Hounds, Golden
Retrievers, Brittany Spaniels, and Labrador Retrievers) were ranked at or below the fourth
decile for both ‘watchdog behavior’ and ‘territorial defense’ in Hart and Hart (1988) survey.
Siberian Huskies ranked in the second decile for ‘watchdog behavior’ but in the sixth decile
454 D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460
Fig. 5. Mean scores (95% confidence intervals) of stranger-, owner- and dog-directed aggression of (A) English
Springer Spaniels and (B) Labrador Retrievers comparing dogs bred for conformation (dark gray) vs. field stock (light
P < 0.05.
for ‘territorial defense’ while English Springer Spaniels showed the opposite pattern (Hart
and Hart, 1988). Bernese Mountain Dogs, Greyhounds and Whippets were not included in the
Hart and Hart (1988) study.
The relatively average C-BARQ scores for stranger-directed aggression found among Pit Bull
Terriers (Fig. 3A) were inconsistent with their universal reputation as a ‘dangerous breed’ and
their reported involvement in dog bite-related fatalities (Sacks et al., 1996). In our survey, nearly
7% of Pit Bull owners indicated that their dogs had bitten or attempted to bite an unfamiliar
person in the recent past, somewhat higher than the overall average (4.7%), while 22% reported
bites directed at other dogs. This pattern is consistent with the view that this breed has been
selectively bred for aggression toward other dogs rather than humans (Lockwood, 1995). It
should be emphasized, however, that while the prevalence of human-directed bites or bite
attempts among Pit Bull Terriers may be only slightly above average, the severity of their attacks
is probably affected by other traits (e.g., the size and strength of the breed, its reputed failure to
give warning signs, and its reported tenacity when attacking) that may also have been selected for
in the development of this ‘‘fighting’’ breed. In contrast, although more than 20% of Dachshund
owners in our study reported bites or attempts to bite against humans, the relatively small size of
this and other highly aggressive breeds (e.g., Chihuahuas) substantially reduces the risks of
4.1.2. Owner-directed aggression (ODA)
In general, scores for ODA were very low and most owners reported no signs of aggression
towards themselves or other members of the household in any context. More than half of cases
involving severe aggression (bites or attempts to bite) were associated with a household member
taking food or other valued objects away from the dog. The low prevalence of ODA in all breeds
makes sense from an evolutionary perspective, since this type of aggression, in contrast to that
directed at strangers and other dogs, has probably never been encouraged by human owners, and
is likely to have been actively selected against.
Breeds that stood out as being rated relatively high (above the median) for aggression toward
household members in our study included the Basset Hound, Beagle, Chihuahua, American
Cocker Spaniel, Dachshund, English Springer Spaniel and Jack Russell Terrier (Figs. 1B and
3B); all breeds in the small to medium size range. Presumably, aggression among larger more
powerful breeds would be more difficult to tolerate or manage. The higher levels of ODA among
English Springer Spaniels in both samples concur with recent published reports of problems with
dominance-type aggression in this breed (Borchelt, 1983; Reisner et al., 1994; Guy et al., 2001a;
Reisner et al., 2005). In their survey of clients sampled from a general veterinary caseload in
Canada, Guy et al. (2001a) reported that English Springer Spaniels were the breed most often
cited to have bitten members of the household (26.8% of owners reported biting). In general,
percentages of dogs reported as having bitten household members by Guy et al. (2001a) were
substantially greater than those reported here (average rate of all breeds combined 13.2% vs.
2%, respectively), a discrepancy that may be partly attributable to the present study’s focus on
aggression only in the recent past.
Aggression directed towards people living in the household is often interpreted as a result of
conflicts related to social dominance (for review, see Lindsay, 2001, pp. 229–272). In support of
this, we found a highly significant correlation between owner-directed aggression and rivalry
among dogs living in the same home. This correlation was independent of aggression toward
unfamiliar people or dogs, suggesting that some canine aggression is specific to those individuals
with whom the dog is familiar and is consistent with the view that aggression towards owners is
D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460 455
sometimes related to social dominance, especially with respect to food or object possession/
4.1.3. Dog-directed aggression (DDA)
In general, we found higher levels of aggression directed toward unfamiliar dogs compared to
unfamiliar people (parts A and C in Figs. 1 and 3); however, this pattern was highly breedspecific.
Dachshunds, for example, showed similar levels of aggression to both dogs and humans
(parts A and C in Figs. 1 and 3) while Akitas, Jack Russell Terriers and Pit Bull Terriers showed
substantially greater aggression toward dogs (Fig. 3A and C).
Our study found significant differences across breeds in displays of aggression toward
unfamiliar dogs and several breeds stood out as being particularly aggressive: Akita, Boxer,
Australian Cattle Dog, German Shepherd, Pit Bull, Chihuahua, Dachshund, English Springer
Spaniel, Jack Russell Terrier and West Highland White Terrier (Figs. 1C and 3C). Six of these
breeds (Akita, Boxer, Chihuahua, Dachshund, German Shepherd, and West Highland White
Terrier) were ranked at the sixth decile or higher for ‘aggression toward other dogs’ in Hart and Hart
(1988) report. English Springer Spaniels ranked in the second decile, and Australian Cattle Dogs,
Pit Bull Terriers and Jack Russell Terriers were not included in the Hart and Hart (1988) survey.
A detailed analysis of a German population of dogs revealed that, among other breeds, Pit Bull
Terriers, German Shepherds, Great Danes and Rottweilers were often the aggressors in inter-dog
conflicts, while Boxers, Cocker Spaniels (presumably English), Dachshunds, Doberman
Pinschers, Poodles, Yorkshire Terriers and West Highland White Terriers were more often the
victims (Roll and Unshelm, 1997). In our study, most of the breeds that Roll and Unshelm (1997)
cited as being aggressors tended to score higher for aggression than fear relative to the population
average (Figs. 2B and 4B).
4.2. Aggression and fear
The present findings point to an interesting balance of aggressive and fearful motivations
underlying the expression of aggressive behavior in the various breeds. While aggression is often
associated with fear in animals (Wingfield et al., 2006, pp. 179–182), this relationship appears to
be stronger in some dog breeds than others. For example, Rottweilers were below average for fear
of strangers but above average for stranger-directed aggression (Fig. 2A). Doberman Pinschers,
Jack Russell Terriers, West Highland White Terriers, Australian Cattle Dogs and German
Shepherds were also more aggressive than fearful towards strangers (Fig. 4A). In contrast,
Dachshunds, Chihuahuas and Yorkshire Terriers were well above average for both aggression
and fear (Figs. 2A and 4A), while Shetland Sheepdogs and Greyhounds tend to be more fearful
than aggressive (Figs. 2A and 4A). These results make intuitive sense given that the more
aggressive than fearful breeds in our study have historically held working roles that require some
degree of assertiveness (protection, herding and hunting) (American Kennel Club, 1992). By
expressing this balance between defensive (fear-mediated) and offensive aggression to widely
different degrees, dog breeds may represent a useful model for studying the underlying causation
of aggressive behavior.
4.3. Conformation vs. field stock
Our results indicate that owner-directed aggression is more pronounced in conformation-bred
English Springer Spaniels compared to field stock dogs, replicating findings in the literature
456 D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460
(Reisner et al., 2005). In contrast, field-bred Labrador Retrievers obtained significantly higher
ODA scores than conformation-bred dogs. The fact that these two breeds showed opposite
patterns confirms Reisner et al.’s (2005) finding that the higher aggression in show-bred English
Springer Spaniels is attributable to a popular sire effect rather than breeding for show per se. A
Swedish study of 31 different breeds found that, in general, breeding for show was associated
with lower levels of aggression, curiosity and playfulness, and with higher levels of fearfulness
(Svartberg, 2006). In contrast, selection for use in field trials was correlated with higher levels of
playfulness and aggression (Svartberg, 2006). Taken together, these findings suggest that canine
aggression has some genetic basis, and that aggressiveness may be selected for either
intentionally or inadvertently by different breeding practices.
While the results of the present study demonstrate striking and consistent variation in
aggression among dog breeds, they shed little light on the underlying sources of this variation in
behavior. Demographic and environmental risk factors for the development of canine aggression
need to be investigated across a variety of breeds so that both generalized and breed-specific
influences can be identified. More empirical data regarding the effects of hormones and neuter
status among the various breeds are also needed. Genetic and environmental factors are likely to
interact to mediate the expression of aggressive behavior during development. Genetic markers
associated with aggressiveness in particular contexts are likely to be identified in the near future
due to the recent sequencing of the dog genome (Lindblad-Toh et al., 2005). Using valid and
reliable measures of canine behavioral phenotypes, such as the C-BARQ, behavioral genetic
studies will further our understanding of how aggressive traits are inherited and mediated by
experiential and environmental factors.
We found large and consistent differences among dog breeds in the prevalence and severity of
aggression directed at different targets (familiar and unfamiliar humans and dogs), and the degree
to which aggression was associated with fear. Reported levels of aggression in some cases are
concerning, with rates of bites or bite attempts ‘‘in the recent past’’ rising as high as 20% toward
strangers and 30% toward unfamiliar dogs in some breeds. In general, the highest rates of humandirected
aggression were found in smaller breeds whose aggression is presumably easier to
tolerate. Differences between lines of distinct breeding stock indicate that the propensity toward
aggressive behavior is at least partially rooted in genetics, although substantial within-breed
variation suggests that other factors (developmental, environmental) play a major part in
determining whether aggressive behavior is expressed in the phenotype. The study also
demonstrates the value of the internet for collecting population-level behavioral data on dogs. In
the future, the use of standardized measures of canine behavioral phenotypes, such as the CBARQ,
by owners and breeders may help to illuminate the causes and reduce the prevalence of
aggression in pet dogs.
The authors thank all the breed club members and dog owners who participated in the surveys
as well as two anonymous reviewers whose comments greatly improved the manuscript.
Development of the C-BARQ was made possible by grants from the Kenneth Scott Charitable
Trust, the Arell Foundation, the University of Pennsylvania Research Foundation, the AKC
Canine Health Foundation, and the Pet Care Trust.
D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460 457
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, at
American Kennel Club, 1992. The Complete Dog Book, eighteenth edition. Howell Book House, New York, pp. 253–
Bamberger, M., Houpt, K.A., 2006. Signalment factors, comorbidity, and trends in behavior diagnoses in dogs: 1644 cases
(1991–2001). J. Am. Vet. Med. Assoc. 229, 1591–1601.
Bartko, J.J., 1966. Intraclass correlation coefficient as a measure of reliability. Psychol. Rep. 19, 3–7.
Beaver, B.V., 1983. Clinical classification of canine aggression. Appl. Anim. Ethol. 10, 35–43.
Blackshaw, J.K., 1991. An overview of types of aggressive behavior in dogs and methods of treatment. Appl. Anim.
Behav. Sci. 30, 351–361.
Borchelt, P., 1983. Aggressive-behavior of dogs kept as companion animals—classification and influence of sex,
reproductive status and breed. Appl. Anim. Ethol. 10, 45–61.
Bradshaw, J.W.S., Goodwin, D., 1998. Determination of behavioural traits of pure-bred dogs using factor analysis and
cluster analysis; a comparison of studies in the USA and UK. Res. Vet. Sci. 66, 73–76.
Carmines, E.G., Zeller, R.A., 1979. Reliability and Validity Assessment. Sage Publications, Thousand Oaks, CA, p. 51.
Centers for Disease Control and Prevention, 2003. Nonfatal dog bite-related injuries treated in hospital emergency
departments—United States, 2001. MMWR 52, 605–628.
Christensen, E., Scarlett, J., Campagna, M., Houpt, K.A., 2007. Aggressive behavior in adopted dogs that passed a
temperament test. Appl. Anim. Behav. Sci. 106, 85–95.
Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences, second edition. Lawrence Earlbaum Associates,
Hillsdale, NJ, pp. 20–26.
Cronbach, L.J., 1951. Coefficient alpha and the internal structure of tests. Psychometrika 16, 297–334.
Gershman, K., Sacks, J., Wright, J., 1994. Which dogs bite—a case-control study of risk-factors. Pediatrics 93, 913–917.
Gosling, S.D., Vazire, S., Srivastava, S., John, O.P., 2004. Should we trust web-based studies? A comparative analysis of
six preconceptions about internet questionnaires. Am. Psychol. 59, 93–104.
Guy, N.C., Luescher, U.A., Dohoo, S.E., Spangler, E., Miller, J.B., Dohoo, I.R., Bate, L.A., 2001a. Demographic and
aggressive characteristics of dogs in a general veterinary caseload. Appl. Anim. Behav. Sci. 74, 15–28.
Guy, N.C., Luescher, U.A., Dohoo, S.E., Spangler, E., Miller, J.B., Dohoo, I.R., Bate, L.A., 2001b. Risk factors for dog
bites to owners in a general veterinary caseload. Appl. Anim. Behav. Sci. 74, 29–42.
Hart, B., Hart, L., 1985. Selecting pet dogs on the basis of cluster-analysis of breed behavior profiles and gender. J. Am.
Vet. Med. Assoc. 186, 1181–1185.
Hart, B.L., Hart, L.A., 1988. The Perfect Puppy: How to Choose Your Dog by its Behavior. W.H. Freeman and Company,
New York, pp. 1–182.
Hsu, Y., Serpell, J.A., 2003. Development and validation of a questionnaire for measuring behavior and temperament
traits in pet dogs. J. Am. Vet. Med. Assoc. 223, 1293–1300.
Kroll, T.L., Houpt, K.A., Erb, H.N., 2004. The use of novel stimuli as indicators of aggressive behavior in dogs. J. Am.
Anim. Hosp. Assoc. 40, 13–19.
Landis, J.R., Koch, G.G., 1977. Measurement of observer agreement for categorical data. Biometrics 33, 159–174.
Liinamo, A., van den Berg, L., Leegwater, P.A.J., Schilder, M.B.H., van Arendonk, J.A.M., van Oost, B.A., 2007. Genetic
variation in aggression-related traits in golden retriever dogs. Appl. Anim. Behav. Sci. 104, 95–106.
Lindblad-Toh, K., Wade, C.M., Mikkelsen, T.S., Karlsson, E.K., Jaffe, D.B., Kamal, M., Clamp, M., Chang, J.L.,
Kulbokas 3rd., E.J., Zody, M.C., Mauceli, E., Xie, X., Breen, M., Wayne, R.K., Ostrander, E.A., Ponting, C.P.,
Galibert, F., Smith, D.R., DeJong, P.J., Kirkness, E., Alvarez, P., Biagi, T., Brockman, W., Butler, J., Chin, C.W., Cook,
A., Cuff, J., Daly, M.J., DeCaprio, D., Gnerre, S., Grabherr, M., Kellis, M., Kleber, M., Bardeleben, C., Goodstadt, L.,
Heger, A., Hitte, C., Kim, L., Koepfli, K.P., Parker, H.G., Pollinger, J.P., Searle, S.M., Sutter, N.B., Thomas, R.,
Webber, C., Baldwin, J., Abebe, A., Abouelleil, A., Aftuck, L., Ait-Zahra, M., Aldredge, T., Allen, N., An, P.,
Anderson, S., Antoine, C., Arachchi, H., Aslam, A., Ayotte, L., Bachantsang, P., Barry, A., Bayul, T., Benamara, M.,
Berlin, A., Bessette, D., Blitshteyn, B., Bloom, T., Blye, J., Boguslavskiy, L., Bonnet, C., Boukhgalter, B., Brown, A.,
Cahill, P., Calixte, N., Camarata, J., Cheshatsang, Y., Chu, J., Citroen, M., Collymore, A., Cooke, P., Dawoe, T., Daza,
R., Decktor, K., DeGray, S., Dhargay, N., Dooley, K., Dooley, K., Dorje, P., Dorjee, K., Dorris, L., Duffey, N., Dupes,
458 D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460
A., Egbiremolen, O., Elong, R., Falk, J., Farina, A., Faro, S., Ferguson, D., Ferreira, P., Fisher, S., FitzGerald, M.,
Foley, K., Foley, C., Franke, A., Friedrich, D., Gage, D., Garber, M., Gearin, G., Giannoukos, G., Goode, T., Goyette,
A., Graham, J., Grandbois, E., Gyaltsen, K., Hafez, N., Hagopian, D., Hagos, B., Hall, J., Healy, C., Hegarty, R.,
Honan, T., Horn, A., Houde, N., Hughes, L., Hunnicutt, L., Husby, M., Jester, B., Jones, C., Kamat, A., Kanga, B.,
Kells, C., Khazanovich, D., Kieu, A.C., Kisner, P., Kumar, M., Lance, K., Landers, T., Lara, M., Lee, W., Leger, J.P.,
Lennon, N., Leuper, L., LeVine, S., Liu, J., Liu, X., Lokyitsang, Y., Lokyitsang, T., Lui, A., Macdonald, J., Major, J.,
Marabella, R., Maru, K., Matthews, C., McDonough, S., Mehta, T., Meldrim, J., Melnikov, A., Meneus, L., Mihalev,
A., Mihova, T., Miller, K., Mittelman, R., Mlenga, V., Mulrain, L., Munson, G., Navidi, A., Naylor, J., Nguyen, T.,
Nguyen, N., Nguyen, C., Nguyen, T., Nicol, R., Norbu, N., Norbu, C., Novod, N., Nyima, T., Olandt, P., O’Neill, B.,
O’Neill, K., Osman, S., Oyono, L., Patti, C., Perrin, D., Phunkhang, P., Pierre, F., Priest, M., Rachupka, A.,
Raghuraman, S., Rameau, R., Ray, V., Raymond, C., Rege, F., Rise, C., Rogers, J., Rogov, P., Sahalie, J., Settipalli, S.,
Sharpe, T., Shea, T., Sheehan, M., Sherpa, N., Shi, J., Shih, D., Sloan, J., Smith, C., Sparrow, T., Stalker, J., StangeThomann,
N., Stavropoulos, S., Stone, C., Stone, S., Sykes, S., Tchuinga, P., Tenzing, P., Tesfaye, S., Thoulutsang, D.,
Thoulutsang, Y., Topham, K., Topping, I., Tsamla, T., Vassiliev, H., Venkataraman, V., Vo, A., Wangchuk, T., Wangdi,
T., Weiand, M., Wilkinson, J., Wilson, A., Yadav, S., Yang, S., Yang, X., Young, G., Yu, Q., Zainoun, J., Zembek, L.,
Zimmer, A., Lander, E.S., 2005. Genome sequence, comparative analysis and haplotype structure of the domestic dog.
Nature 438, 803–819.
Lindsay, S.R., 2001. Handbook of Applied Dog Behavior and Training Volume Two: Etiology and Assessment of
Behavior Problems. Iowa State University Press, Ames, IA, pp. 161–272.
Lockwood, R., 1995. The ethology and epidemiology of canine aggression. In: Serpell, J.A. (Ed.), The Domestic Dog: Its
Evolution, Behavior and Interactions with People. Cambridge University Press, Cambridge, pp. 131–138.
Nederhof, A.J., 1985. Methods of coping with social desirability bias: a review. Eur. J. Soc. Psychol. 15, 263–280.
Netto, W.J., Planta, D.J.U., 1997. Behavioral testing for aggression in the domestic dog. Appl. Anim. Behav. Sci. 52, 243–
Nunnally, J.C., 1978. Psychometric Theory. McGraw-Hill, New York, p. 245.
Notari, L., Goodwin, D., 2007. A survey of behavioural characteristics of pure-bred dogs in Italy. Appl. Anim. Behav. Sci.
Planta, J.U.D., De Meester, R.H.W.M., 2007. Validity of the Socially Acceptable Behavior (SAB) test as a measure of
aggression in dogs towards non-familiar humans. Vlaams Diergeneeskundig Tijdschrift 76, 359–368.
Reisner, I., Erb, H., Houpt, K., 1994. Risk factors for behavior-related euthanasia among dominant-aggressive dogs—110
cases (1989–1992). J. Am. Vet. Med. Assoc. 205, 855–863.
Reisner, I.R., Houpt, K.A., Shofer, F.S., 2005. National survey of owner-directed aggression in English Springer Spaniels.
J. Am. Vet. Med. Assoc. 227, 1594–1603.
Roll, A., Unshelm, J., 1997. Aggressive conflicts amongst dogs and factors affecting them. Appl. Anim. Behav. Sci. 52,
Sacks, J.J., Lockwood, R., Hornreich, J., Sattin, R.W., 1996. Fatal dog attacks, 1989–1994. Pediatrics 97, 891–895.
Salman, M.D., New Jr., J.G., Scarlett, J.M., Kass, P.H., Ruch-Gallie, R., Hetts, S., 1998. Human and animal factors related
to relinquishment of dogs and cats in 12 selected animal shelters in the United States. J. Appl. Anim. Welfare Sci. 1,
Serpell, J.A., Hsu, Y., 2005. Effects of breed, sex, and neuter status on trainability in dogs. Anthrozoos 18, 196–207.
Serpell, J.A., Jagoe, J.A., 1995. Early experience and the development of behaviour. In: Serpell, J.A. (Ed.), The
Domestic Dog: Its Evolution, Behaviour and Interactions with People. Cambridge University Press, Cambridge,
Shrout, P.E., Fleiss, J.L., 1979. Intraclass correlations: uses in assessing rater reliability. Psychol. Bull. 86, 420–428.
Svartberg, K., 2005. A comparison of behavior in test and in everyday life: evidence of three consistent boldness-related
personality traits in dogs. Appl. Anim. Behav. Sci. 91, 103–128.
Svartberg, K., 2006. Breed-typical behavior in dogs—historical remnants or recent constructs? Appl. Anim. Behav. Sci.
Takeuchi, Y., Mori, Y., 2006. A comparison of the behavioral profiles of purebred dogs in Japan to profiles of those in the
United States and the United Kingdom. J. Vet. Med. Sci. 68, 789–796.
van den Berg, L., Schilder, M.B.H., Knol, B.W., 2003. Behavior genetics of canine aggression: behavioral phenotyping of
golden retrievers by means of an aggression test. Behav. Genet. 33, 469–483.
van der Borg, J.A.M., Netto, W.J., Planta, D.J.U., 1991. Behavioural testing of dogs in animal shelters to predict problem
behaviour. Appl. Anim. Behav. Sci. 32, 237–251.
Weiss, H.B., Friedman, D.I., Coben, J.H., 1998. Incidence of dog bite injuries treated in emergency departments. J. Am.
Med. Assoc. 279, 51–53.
D.L. Duffy et al. / Applied Animal Behaviour Science 114 (2008) 441–460 459
Wingfield, J.C., Moore, I.T., Goymann, W., Wacker, D.W., Sperry, T., 2006. Contexts and ethology of vertebrate
aggression: implications for the evolution of hormone-behavior interactions. In: Nelson, R.J. (Ed.), Biology of
Aggression. Oxford University Press, New York, pp. 179–182.
Wright, J., Nesselrote, M., 1987. Classification of behavior problems in dogs—distributions of age, breed, sex and
reproductive status. Appl. Anim. Behav. Sci. 19, 169–178.
Wright, J.C., 1991. Canine aggression toward people. In: Marder, A.R., Voith, V. (Eds.), Veterinary Clinics of North
America Small Animal Practice: Advances in Companion Animal Behavior., vol. 21. W.B. Saunders Company,
Philadelphia, PA, pp. 299–314.