Content uploaded by James A Serpell
Author content
All content in this area was uploaded by James A Serpell
Content may be subject to copyright.
B
ehavioral problems are a leading cause of relin-
quishment of dogs to shelters; in their decision to
relinquish, 47% of people cite behavioral reasons as a
contributing factor.
1-3
The most common behavioral
reasons for relinquishment include aggression toward
people, aggression toward animals, escaping, destruc-
tion, disobedience, house soiling, and excessive bark-
ing.
4
Although increasing the number of dogs adopted
from shelters can potentially improve welfare and
decrease the euthanasia rate among relinquished dogs,
without valid behavioral screening and assessment
tools, shelter personnel may be unknowingly allowing
dogs to be adopted that may be unsuitable for or pose
a public health risk to the adopting family.
Many shelters rely on observation-based behavior
evaluation tests performed by shelter personnel as a
method of assessing a dog’s behavioral profile and
potential public health risk.
5
However, these tests are
often either of uncertain or unknown reliability and
validity
6
or are too extensive to replicate in most shel-
ter environments.
7,8
Because behavior is assessed in a
novel and potentially stressful environment, often in 1
brief evaluation, there is a possibility that 1 or more
behavioral problems that could become apparent in the
adoptive home may not be detected.
An alternative behavioral assessment technique is
to obtain historical information about pets. An inter-
view with or a questionnaire completed by a person
who is relinquishing a dog has the potential to provide
information about the dog’s behavior in its former
home.
9
However, an owner may be tempted to be less
than candid about serious problem behaviors, such as
human-directed aggression, to increase the likelihood
that the dog will be adopted.
10
This possibility would
reduce the usefulness of data collected via a question-
naire or an interview; additionally, the information
obtained may be misleading and potentially dangerous.
Accordingly, the primary purpose of the study reported
here was to evaluate a behavioral questionnaire in ani-
mal shelters for the presence of biased results by inves-
tigating whether owners who were relinquishing dogs
would be inclined to provide more candid responses to
the questionnaire if they believed that this information
was to be held confidential from shelter staff members
(who owners might presume would determine their
dog’s outcome). Specifically, the aim was to test the null
hypothesis that owners who were informed at the out-
set that their questionnaire responses would be shared
with and used by shelter staff to make adoption deci-
JAVMA, Vol 227, No. 11, December 1, 2005 Scientific Reports: Original Study 1755
SMALL ANIMALS
Evaluation of a behavioral assessment questionnaire
for use in the characterization of behavioral problems
of dogs relinquished to animal shelters
Sheila A. Segurson, DVM; James A. Serpell, PhD; Benjamin L. Hart, DVM, PhD, DACVB
Objective—To evaluate a behavioral intake question-
naire in animal shelters for the presence of biased
results and assess its use in the characterization of
behavioral problems of dogs relinquished to shelters.
Design—Cross-sectional study.
Animals—54 dogs being relinquished to a shelter
and 784 dogs belonging to veterinary clients.
Procedure—Owners who were relinquishing their
dogs and agreed to complete the behavioral ques-
tionnaire were alternately assigned to 1 of 2 groups;
participants were aware that information provided
would be confidential or nonconfidential (ie, likely
used for adoption purposes). Data from confidential
and nonconfidential information groups were com-
pared, and the former were compared with data (col-
lected via the questionnaire) regarding a population of
client-owned dogs.
Results—Analyses revealed significant differences in
2 areas of reported problem behavior between the
confidential and nonconfidential information groups:
owner-directed aggression and stranger-directed fear.
Compared with client-owned–group data, significant-
ly more relinquished shelter dogs in the confidential
information group were reported to have owner-
directed aggression, stranger-directed aggression,
dog-directed aggression or fear, stranger-directed
fear, nonsocial fear, and separation-related behaviors.
Conclusions and Clinical Relevance—Among per-
sons relinquishing dogs to a shelter, those who
believed questionnaire responses were confidential
reported owner-directed aggression and fear of
strangers in their pets more frequently than relinquish-
ers who believed responses were nonconfidential.
Confidentiality had no apparent effect on the reporting
of other assessed behavioral problems. Results sug-
gest that behavioral questionnaires may sometimes
provide inaccurate information in a shelter setting, but
the information may still be useful when evaluating
behavior of relinquished dogs. (
J Am Vet Med Assoc
2005;227:1755–1761)
From the Center for Companion Animal Health, School of Veterinary Medicine (Segurson), and the Behavior Service, Veterinary Medical
Teaching Hospital (Hart), University of California, Davis, CA 95616; and the Department of Clinical Studies, School of Veterinary Medicine,
University of Pennsylvania, Philadelphia, PA 19104-6010 (Serpell).
Supported by Maddie’s Fund and the University of California, Davis, Center for Companion Animal Health.
Presented at the House Officer Seminar Day, University of California, Davis, Calif, April 2005 and the 5th International Veterinary Behavior
Meeting, Minneapolis, July 2005.
The authors thank Dr. Melissa Bain for technical assistance and Dr. Phillip Kass for statistical assistance.
Address correspondence to Dr. Segurson.
05-01-0034.qxp 11/11/2005 12:53 PM Page 1755
sions would score their dogs no differently from owners
who were promised confidentiality in this regard.
Risk factors for relinquishment of pets to a shelter
identified in previous studies
2,3
include behavioral
problems such as aggression toward people, aggression
toward animals, and destruction. These studies provid-
ed useful information but did not evaluate specific
behavioral diagnoses that could be used to more accu-
rately characterize the nature of behavioral problems in
shelter animals and assess potential public health risks.
Therefore, a second purpose of the study reported here
was to assess the use of a behavioral questionnaire in
the characterization of behavioral problems of dogs
relinquished to shelters by statistical comparison of
questionnaire-derived behavioral profiles of shelter
dogs with profiles reported via the same questionnaire
by owners of dogs not being relinquished to shelters.
Results were expected to provide information about
the types and prevalence of problem behaviors in shel-
ter dogs, compared with a population of client-owned
dogs that were not being relinquished to a shelter.
Materials and Methods
Questionnaire—The 103-item version of the Canine
Behavioral Assessment and Research Questionnaire (C-
BARQ)
a
was used to obtain behavioral information from the
owners of relinquished dogs. The development of the ques-
tionnaire (formerly known as PennBARQ) has been
described in detail.
11
Briefly, the questionnaire consists of a
series of 5-point behavioral rating scales and is designed to be
used to screen dogs for the presence and severity of problem
behaviors. A score of 0 represents the absence of the behav-
ior, with scores of 1 through 4 representing increasing sever-
ity of the problem. The questionnaire measures 11 distinct
temperament factors or traits, 10 of which have been shown
to possess adequate internal consistency (Cronbach α
> 0.7), and 7 of which were successfully validated by com-
parison with a population of dogs that had clinically diag-
nosed behavioral problems. Only these 7 validated factors
were used in the present study. The factors were labeled and
defined as follows: stranger-directed aggression (9 question-
naire items related to a dog’s tendency to respond aggressive-
ly to strangers approaching or invading its or its owner’s
personal space, territory, or home range), owner-directed
aggression (8 items related to a tendency to respond aggres-
sively to the owner or other members of the household when
stared at, stepped over, scolded, bathed or groomed,
approached while in possession of food or objects, or when
food or objects are taken away), stranger-directed fear (3
items related to a tendency to respond fearfully when
approached directly by strangers), nonsocial fear (6 items
related to a tendency to react fearfully to sudden or loud
noises and unfamiliar objects and situations), dog-directed
fear or aggression (5 items related to a tendency to respond
fearfully or aggressively when approached directly by unfa-
miliar dogs), separation-related behavior (8 items related to a
tendency to vocalize or engage in destructive behavior when
separated from the owner, which was accompanied or pre-
ceded by behavioral and autonomic signs of anxiety includ-
ing restlessness, loss of appetite, trembling, and excessive
salivation), and attachment or attention-seeking behavior (6
items related to a tendency to maintain close proximity to the
owner or other members of the household, to solicit affection
or attention, and to become agitated when the owner gives
attention to third parties).
A dog’s score for each factor was calculated as the mean
of its scores on each questionnaire item included in that fac-
tor. Thus, a score of 1.200 for stranger-directed aggression
would reflect the mean score of the 9 questions assigned to
that factor for that dog. If a particular respondent did not
answer an item included in factor calculation, that respon-
dent was eliminated from the calculation of that factor, and
the sample size for statistical purposes was adjusted accord-
ingly. Thus, the sample size for some of the categories was
different from others for the same group of respondents.
To determine whether people provided biased responses to
the questionnaire on the basis of the perceived impact of the data
on their pet’s chances of adoption, the questionnaire was intro-
duced by 2 different preambles. The questionnaire for 1 group
(the nonconfidential information group) stated that the data col-
lected would be used by shelter personnel to match the dog to
an appropriate home; the questionnaire for the other group (the
confidential information group) stated that the information pro-
vided by respondents would be held confidential.
Participants—To collect data via nonconfidential and
confidential questionnaires, people who were relinquishing
dogs to either the Sacramento County Department of Animal
Care and Regulation or the Sacramento Society for
Prevention of Cruelty to Animals were solicited. Trained
assistants visited the 2 shelters several days per week over a
2-month period; the cumulative time spent at both shelters
was 240 hours. The research assistants requested voluntary
participation from people relinquishing their dogs. A person
was excluded from participation in the study if the dog being
relinquished was < 4 months old, the dog had been owned
for ≤ 3 months, the dog was being relinquished for euthana-
sia, or the person had poor English skills. Subjects were alter-
nately assigned to the confidential or nonconfidential infor-
mation groups. Thus, it was assumed that there would be no
appreciable differences in actual problem behaviors of dogs
in the 2 groups. Because we hypothesized that persons
assigned to the nonconfidential information group might
complete the questionnaire more quickly than persons
assigned to the confidential information group, time to com-
plete the questionnaire was recorded for each respondent.
For comparison with scores of dogs being relinquished,
scores were collected for a population of owned (nonrelin-
quished) dogs; data were collected from 784 clients of the
Veterinary Hospital of the University of Pennsylvania who
had visited the hospital in the preceding 3 years with a dog
for which the behavioral assessment questionnaire was com-
pleted (client-owned group).
11
Data from client-owned dogs
that were < 1 or > 7 years old were excluded, along with data
from dogs that had severe or chronic health problems and
dogs that had been examined because of a behavioral prob-
lem. Clients of the veterinary teaching hospital were asked to
check a box if their pet had any severe or chronic health
problems or had been examined because of a behavioral
problem; data from client-owned dogs were excluded from
the study on the basis of this answer. For comparison of
client-owned–dog scores with scores for relinquished shelter
dogs, only the questionnaires from the confidential informa-
tion group were included on the assumption that the infor-
mation from that group would be a more accurate reflection
of the behavior of dogs being relinquished to shelters.
Statistical analysis—Responses of the confidential
information group were compared with those of the noncon-
fidential information group in 2 ways. First, the distributions
of zero scores (which could reflect either the absence of
behavioral problems or a deliberate attempt by an owner to
misrepresent the dog’s prevalence of behavioral problems)
and nonzero scores were compared by use of a Fisher exact
test.
12
Second, a Mann-Whitney U test was used to compare
the distribution of questionnaire scores on each of the behav-
ioral factors between the 2 groups.
b
1756 Scientific Reports: Original Study JAVMA, Vol 227, No. 11, December 1, 2005
SMALL ANIMALS
05-01-0034.qxp 11/11/2005 12:53 PM Page 1756
A simplified scoring system was used to
reflect an overall assessment of the presence
or absence of behavioral problems. Mean
score data were divided into 2 categories: ≤ 1
(low to no evidence of behavioral problems)
and > 1 (moderate to severe behavioral prob-
lems). χ
2
Tests were used to compare the dis-
tribution of these categories between the shel-
ter-derived confidential information group
and the client-owned group.
12
Attachment or
attention-seeking scores were not included in
this analysis because the distribution of scores
for this factor did not lend itself to this type of
binary categorization.
Prevalence odds ratios and 95% confidence
intervals were calculated to compare the preva-
lence odds of behavioral problems in the confi-
dential group of relinquished dogs versus
client-owned dogs.
13
χ
2
Tests were used to com-
pare the distribution of neuter status (neutered
vs sexually intact dogs), sex of dog, and sex of
person completing the questionnaire between
the 2 groups.
12
Age of the dog, duration of own-
ership, and amount of time to complete the
questionnaire were compared by use of a
Student t test.
13
A value of P < 0.05 was consid-
ered significant.
Results
Comparison of data collected from confidential
and nonconfidential information groups—Of 204 peo-
ple that relinquished dogs during observation hours,
100 (49%) met the inclusion criteria; of these 100 indi-
viduals, 54 completed the questionnaire (27 were
assigned to the confidential information group, and 27
were assigned to the nonconfidential information
group). The most common reasons for exclusion were
relinquishment of a stray dog or a dog for euthanasia.
The time taken to complete the questionnaire by partic-
ipants in the confidential information group appeared to
be greater than that taken by participants in the non-
confidential information group (mean time to comple-
tion, 12.3 and 10.9 minutes, respectively); however, this
JAVMA, Vol 227, No. 11, December 1, 2005 Scientific Reports: Original Study 1757
SMALL ANIMALS
Confidential Nonconfidential
Variable information group* (% [n]) information group* (% [n])
Characteristics of relinquishers
Men 63 (17/27) 63 (17/27)
Women 37 (10/27) 33 (9/27)
Couples 0 (0/27) 4 (1/27)
Characteristics of relinquished dogs
Sex
Male 48 (13/27) 67 (18/27)
Female 37 (10/27) 33 (9/27)
Unknown 15 (4/27) 0 (0/27)
Reproductive (neuter) status
Neutered 33 (9/27) 30 (8/27)
Sexually intact 37 (10/27) 37 (10/27)
Unknown 30 (8/27) 33 (9/27)
Age
ⱕ 6 mo 0 (0/27) 7 (2/27)
⬎ 6 to 12 mo 37 (10/27) 22 (6/27)
⬎ 12 to 24 mo 19 (5/27) 22 (6/27)
⬎ 24 to 48 mo 15 (4/27) 15 (4/27)
⬎ 48 mo 30 (8/27) 33 (9/27)
Duration of ownership by relinquisher
⬍ 6 mo 15 (4/27) 11 (3/27)
6 mo to 1 y 26 (7/27) 37 (10/27)
⬎ 1 to 2 y 15 (4/27) 22 (6/27)
⬎ 2 to 5 y 33 (9/27) 15 (4/27)
⬎ 5 y 11 (3/27) 15 (4/27)
*The questionnaire for the confidential information group stated that information provided by respondents
would be held confidential; the questionnaire for the nonconfidential information group stated that the data
collected would be used by shelter personnel for adoption purposes. Mean time required to complete ques-
tionnaire was 12.2 minutes in the confidential information group and 10.9 minutes in the nonconfidential infor-
mation group.
Table 1—Demographic information obtained from behavioral assessment questionnaires completed
for dogs by persons who were relinquishing those dogs to shelters.
Figure 1—Percentage of dogs assigned a mean score of 0 for various behavioral
categories by persons who completed a behavioral assessment questionnaire as
the dogs were relinquished to a shelter. A score of 0 represented the absence of
the behavior, and scores of 1 through 4 represented increasing severity of the prob-
lem. Respondents were informed that their responses would be kept confidential
(open bars; n = 27) or that the responses would be nonconfidential and used by
shelter personnel to match the dog to an appropriate adoptive home (solid bars; 27).
*Value for the nonconfidential information group differs significantly (
P
< 0.05) from
that of the confidential information group in this behavioral category.
05-01-0034.qxp 11/11/2005 12:53 PM Page 1757
difference was not significant. The number of dogs relin-
quished by men, women, or couples did not differ sig-
nificantly. The characterization of dogs in the 2 groups
with regard to age, neuter status, duration of ownership
by the relinquisher, and breed was comparable between
groups (Table 1).
Statistical analyses revealed significant differences
in 2 variables between the confidential and nonconfi-
dential information groups. With regard to aggression
toward owners and members of the family, the score for
dogs relinquished by respondents in the confidential
information group (median score, 0.500; mean score,
0.683; n = 25) was significantly (P = 0.001) greater
than the score for dogs relinquished by respondents in
the nonconfidential information group (median score,
0.000; mean score, 0.234; 23). With regard to stranger-
directed fear, the score for dogs relinquished by
respondents in the confidential information group
(median score, 1.000; mean score, 1.235; n = 27) was
significantly (P = 0.029) greater than the score for dogs
relinquished by respondents in the nonconfidential
information group (median score, 0.084; mean score,
0.646; 24). Analyses of the remaining 5 factors
(stranger-directed aggression, dog-directed aggression
or fear, nonsocial fear, separation-related problems,
and attachment or attention-seeking behavior) did not
reveal significant differences in questionnaire scores
between the 2 groups. χ
2
Analysis of the frequency of
zero scores achieved similar measures of significance
(Figure 1).
Association of behavioral problems
with relinquishment to a shelter—Because
the confidential and nonconfidential infor-
mation groups differed with regard to 2
assessment categories and these differences
were considered likely reflections of bias,
data from only the confidential information
group were used for comparison with data
collected from the questionnaires complet-
ed for 784 client-owned dogs. There was no
significant difference in age, sex, or neuter
status between shelter dogs in the confi-
dential information group and the client-
owned dogs.
χ
2
Analysis of the frequencies of
behavioral factor scores of ≤ 1 and > 1
from the confidential information group,
compared with frequencies among the
client-owned group, revealed 6 behavioral
factors that were associated with signifi-
cant differences in scores between groups
(Figure 2). Compared with client-owned
dogs, dogs being relinquished to shelters
were more likely to have scores ≥ 1 for
owner-directed aggression (P < 0.001),
stranger-directed aggression (P = 0.001),
stranger-directed fear (P < 0.001), aggres-
sion or fear toward dogs (P = 0.001),
nonsocial fear (P = 0.003), and separation-
related behavior (P < 0.001; Table 2).
1758 Scientific Reports: Original Study JAVMA, Vol 227, No. 11, December 1, 2005
SMALL ANIMALS
Figure 2—Percentage of dogs that were being relinquished (confidential informa-
tion group, solid bars; n = 27) and client-owned dogs (open bars; 784) for which
questionnaire respondents assigned a score ≥ 1 (ie, moderate to severe problem)
for 6 behavioral categories. *Value for relinquished dogs differs significantly (
P
<
0.001 to 0.003) from that of the client-owned dogs in this behavioral category.
No. of dogs in
confidential No. of client- Odds 95% Confidence
Behavioral category information group* owned dogs * ratio† interval
Owner-directed aggression 7/25 24/695 10.87 4.15–28.49
Stranger-directed aggression 9/25 131/606 3.63 1.64–8.01
Stranger-directed fear 10/27 91/739 4.19 1.86–9.42
Dog-directed aggression or fear 13/24 138/587 3.85 1.68–8.78
Nonsocial fear 12/24 152/646 3.25 1.42–7.38
Separation-related behavior 11/25 93/715 5.25 2.31–11.92
*Expressed as number of dogs with problem behavior per number of dogs for which data were collected
regarding the behavioral category. †Odds ratio is expressed as the odds of a dog in the confidential informa-
tion group with a mean score ⬎ 1 for the behavioral category, compared with the client-owned–dog group.
Table 2—Prevalence odds ratios and 95% confidence intervals for behavioral problems among dogs
that were being relinquished to a shelter (as indicated by responses to a behavioral assessment ques-
tionnaire completed by relinquishers who had been told that the information provided would remain
confidential), compared with problems among client-owned dogs (as indicated by responses of vet-
erinary clients of a referral hospital who completed the same questionnaire).
05-01-0034.qxp 11/11/2005 12:53 PM Page 1758
Overall, the most frequently reported behavioral
problem for the confidential information group of shel-
ter dogs was dog-directed aggression or fear (13/24
[54%] dogs); the least reported behavioral problem
was owner-directed aggression (7/25 [28%] dogs).
Among client-owned dogs, dog-directed aggression or
fear (138/587 [24%] dogs) and nonsocial fear (152/646
[24%] dogs) were most prevalent; owner-directed
aggression was least frequently reported (24/695 [4%]
dogs; Figure 2).
Discussion
As one of its objectives, the present study was
designed to examine the hypothesis that persons relin-
quishing dogs to a shelter would be more inclined to
provide misleading or biased responses to some items
of a behavioral questionnaire if they believed that this
information would be used by shelter staff to deter-
mine their dog’s outcome. This hypothesis was exam-
ined by comparing behavioral scores obtained from
evaluation of a validated 103-item questionnaire. The
questionnaire for 1 group stated that the results would
be used by shelter personnel to match the dog to an
appropriate home, and the questionnaire for the other
group stated that the information would be held confi-
dential. Between the confidential and nonconfidential
information groups, there were no significant differ-
ences in reported scores for stranger-directed aggres-
sion, dog-directed aggression or fear, nonsocial fear,
separation-related behavior, and attachment or atten-
tion-seeking behavior. Assuming that the actual behav-
ior of dogs in the 2 groups was comparable, these
results suggest that, for some behavioral categories,
people relinquishing dogs to shelters will provide use-
ful answers on intake questionnaires. However, there
were significant differences in scores for owner-direct-
ed aggression and stranger-directed fear between the
confidential and nonconfidential information groups.
This raises concern because aggression toward house-
hold members is a major cause of dog bites in the
United States.
14-17
Because of the seriousness of this
problem, this misrepresentation by persons relinquish-
ing dogs is a major public health concern.
In several studies,
2,4,18,19
aggression toward humans
has been identified as the most common behavioral
cause of relinquishment or return of adopted dogs to
shelters. It is understandable that people adopting a dog
from a shelter do not desire a dog that could behave
aggressively toward them or family members. Because
most people who are relinquishing their dog to a shel-
ter hope that someone else will adopt it,
20
it is logical to
conclude that comments on aggressive behavior toward
humans (especially those within the household) are
most likely to be influenced by whether the question-
naire information was confidential or nonconfidential.
Interestingly, there was no significant difference in
scores for stranger-directed aggression between the con-
fidential and nonconfidential information groups.
Perhaps only behaviors considered to be the most seri-
ous to the relinquishers were reflected as significant
differences. This may reflect a perception that stranger-
directed aggression is less likely to influence adoptabil-
ity or that such behavior may even be desirable.
Because results of the present study provide evi-
dence of misrepresentation on some aspects of the
behavioral assessment questionnaire, it is unknown
whether people who report no behavioral problems by
assigning a score of 0 to their dog on shelter intake
questionnaires truly have a problem-free dog or are
providing biased answers in an effort to increase their
pet’s likelihood of adoption. The significantly higher
frequency of zero scores for some behavioral categories
in the nonconfidential information group suggests that
some of these zero scores probably reflect bias by the
relinquishers. Shelter personnel should be aware that
completed intake forms that include no evidence of a
particular behavioral problem in dogs may sometimes
be inaccurate. Use of a validated questionnaire, such as
that used in the study of this report, can help shelter
staff to focus detailed behavioral evaluations on dogs
for which relinquishers provide ambiguous informa-
tion or report no behavioral problems. Further evalua-
tion of this issue is warranted to determine whether
different approaches to the presentation and adminis-
tration of a questionnaire can address this problem.
Although the results of the present study question
the reliability of some items included in a behavioral
assessment questionnaire when it is completed at the
time of relinquishment of a dog, it is important to note
that relinquishers who knew that the questionnaire
data may affect their dog’s outcome still frequently dis-
closed behavioral concerns about the pet. Thus,
despite its limitations in a shelter setting, the question-
naire used in the study of this report can help shelter
personnel to identify dogs with behavioral problems,
thus allowing intervention to reduce the risk of aggres-
sive events in the future and decreasing the likelihood
that the adopting family will be dissatisfied with its
new dog. The questionnaire used in the present study
provides important information about the pet that is
being relinquished and takes approximately 12 min-
utes to complete. The value of the data collected (as
determined in our study) should help shelter staff to
justify the inconvenience of questionnaire completion
to people who are relinquishing dogs to shelters.
The results of the comparison between behavioral
assessment scores assigned to relinquished dogs and
client-owned dogs indicated that dogs that were being
relinquished to a shelter were more likely to be
described as having moderate to severe levels of dog-
directed aggression or fear, stranger-directed aggres-
sion, stranger-directed fear, owner-directed aggression,
nonsocial fear, and separation-related behaviors. These
results are consistent with previous reports
2-5
of an
increased risk of relinquishment to shelters for dogs
with behavioral problems, compared with dogs that
were not being relinquished to shelters.
It is noteworthy that results of our study reflect a
change from findings of previous studies,
3,9
which indi-
cated that sexually intact dogs were at higher risk for
relinquishment to a shelter than neutered dogs. The
percentage of sexually intact dogs among client-owned
(nonrelinquished) dogs in the present study was simi-
lar to that determined in previous studies; however,
there were fewer sexually intact dogs in the group
being relinquished to a shelter. The reason for this dif-
JAVMA, Vol 227, No. 11, December 1, 2005 Scientific Reports: Original Study 1759
SMALL ANIMALS
05-01-0034.qxp 11/11/2005 12:53 PM Page 1759
ference is unknown; it is possible that as a result of an
increase in spay-neuter campaigns, the number of sex-
ually intact dogs being relinquished to shelters has
gradually declined in proportion to relinquished
neutered dogs, compared with the relative proportion
determined in studies completed several years earlier.
Although our data indicated significant differences
in behavioral assessment scores between relinquished
and client-owned dogs, it should be noted that the
groups compared during this part of the study repre-
sented different populations in different communities,
and this may account for some or all differences. Clients
that visit a referral veterinary hospital may have dogs
with different behavioral profiles, compared with the
dogs of people who relinquish pets to shelters. Also,
behavioral problems among dogs may differ as a func-
tion of the home setting (eg, urban vs suburban) and
geographic region (eg, California vs Pennsylvania).
The magnitude of the differences in prevalence of
behavioral problems between relinquished dogs and
client-owned dogs raises several considerations.
Categories of behavior that would logically affect the
ease with which a person could keep a pet, namely
owner-directed aggression, dog-directed aggression or
fear, nonsocial fear, stranger-directed aggression,
stranger-directed fear, and separation-related behav-
iors, had the highest prevalence among dogs that were
being relinquished to shelters. The prevalence of seri-
ous behavioral problems in relinquished dogs contrasts
with attestations by shelter personnel that relinquish-
ment reflects the fact that owners do not have enough
time or space for a dog as well as unrealistic expecta-
tions and knowledge of normal dog needs and behav-
ior.
9,21
It appears likely that serious behavioral problems
are often the primary explanation for the inability of
the relinquisher to keep the dog and that a perceived
lack of time or space is not the central issue. This high-
lights the need for primary care veterinarians to pro-
vide preventive behavioral health information for their
clients.
22
Studies
23,24
have revealed that dogs obtained from
shelters have a higher risk of separation anxiety disor-
ders than dogs obtained from other sources. This find-
ing has raised the question of whether this increased
risk is the result of an increased rate of relinquishment
of dogs with separation anxiety disorders or whether
relinquished dogs develop separation anxiety disorders
because of factors in the shelter such as environmental
stress and loss of a primary attachment figure. In the
present study, the significantly greater prevalence of
separation-related behaviors among dogs that were
being relinquished, compared with client-owned dogs,
suggests that separation anxiety disorders in dogs
adopted from shelters are often a preexisting condition.
Because aggression toward people is a serious pub-
lic health concern, the value of specific categorization
of aggression through use of the behavioral assessment
questionnaire, such as owner-directed and stranger-
directed aggression, is considerable. Whether the
aggression reflects primarily genetic or environmental
influences, or a combination thereof, an emphasis
must be placed on developing reliable methods of
behavioral assessment for relinquished dogs to protect
public safety. With a clear perspective of the problems
that result in relinquishment, measures for identifying
problems and for problem prevention and resolution
can be developed, thereby reducing the likelihood that
dogs will be relinquished to shelters because of behav-
ioral problems and increasing the likelihood that relin-
quished dogs will be rehomed.
The results of the present study indicate that even
a well-designed and validated behavioral assessment
questionnaire may not always provide accurate infor-
mation in a noncontrolled shelter setting. Persons
relinquishing dogs to a shelter who knew that their
responses to a questionnaire were confidential were
more likely to report owner-directed aggression and
fear of strangers than relinquishers who believed that
shelter staff would assess and use their responses.
However, confidentiality had no apparent effect on the
reporting of other types of behavioral problems,
including stranger-directed aggression, nonsocial fear,
separation-related behavior, dog-directed aggression or
fear, and attachment or attention-seeking behavior.
These findings suggest that behavioral assessment
questionnaires may sometimes provide inaccurate
information when administered to persons relinquish-
ing dogs in a shelter setting but that this information
may still be useful for the evaluation of the behavior of
relinquished dogs.
a. Copies of the C-BARQ questionnaire are available on request
from Dr. James A. Serpell.
b. Minitab 13 statistical software, Minitab Inc, State College, Pa.
References
1. Salman MD, Scarlett JM, Kass PH, et al. Human and animal
factors related to the relinquishment of dogs and cats in 12 selected
animal shelters in the United States. J Appl Anim Welf Sci 1998;1:
207–226.
2. Miller DM, Stats SR, Partlo BS, et al. Factors associated with
the decision to surrender a pet to an animal shelter. J Am Vet Med Assoc
1996;209:738–742.
3. Patronek GJ, Glickman LT, Beck AM, et al. Risk factors for
relinquishment of dogs to an animal shelter. J Am Vet Med Assoc
1996;209:572–581.
4. Salman MD, Hutchison J, Ruch-Gallie R, et al. Behavioral
reasons for relinquishment of dogs and cats to 12 shelters. J Appl
Anim Welf Sci 2000;3:93–106.
5. Reid P, Goldman J, Zawistowski S. Animal shelter behavior
programs. In: Miller L, Zawistowski S, eds. Shelter medicine for vet-
erinarians and staff. Ames, Iowa: Blackwell Publishing Professional,
2004;317–331.
6. Sternberg S. Great dog adoptions: a guide for shelters.
Alameda, Calif: Latham Foundation, 2002;9–28.
7. van der Borg JAM, Netto WJ, Planta DJU. Behavioural test-
ing of dogs in animal shelters. Appl Anim Behav Sci 1991;32:
237–251.
8. Netto WJ, Planta DJU. Behavioural testing for aggression in
the domestic dog. Appl Anim Behav Sci 1997;52:243–263.
9. New JC, Salman MD, King M, et al. Characteristics of shel-
ter-relinquished animals and their owners compared with animals
and their owners in U.S. pet-owning households. J Appl Anim Welf
Sci 2000;3:181–200.
10. Posage JM, Bartlett PC, Thomas DK. Determining factors
for successful adoption of dogs from an animal shelter. J Am Vet Med
Assoc 1998;213:478–482.
11. Hsu Y, Serpell JA. Development and validation of a ques-
tionnaire for measuring behavior and temperament traits in pet dogs.
J Am Vet Med Assoc 2003;223:1293–1300.
12. Siegal S, Castellan NJ. Nonparametric statistics for the behav-
ioral sciences. 2nd ed. Boston: McGraw-Hill Book Co, 1988;103–124.
1760 Scientific Reports: Original Study JAVMA, Vol 227, No. 11, December 1, 2005
SMALL ANIMALS
05-01-0034.qxp 11/11/2005 12:53 PM Page 1760
13. Howell DC. Statistical methods for psychology. 5th ed. Pacific
Grove, Calif: Thomson Learning Inc, 2002;165–167, 198–207.
14. Overall KL, Love M. Dog bites to humans—demography, epi-
demiology, injury, and risk. J Am Vet Med Assoc 2001;218:1923–1934.
15. Guy NC, Luescher UA, Dohoo SE, et al. Demographic and
aggressive characteristics of dogs in a general veterinary caseload.
Appl Anim Behav Sci 2001;74:15–28.
16. Guy NC, Luescher UA, Dohoo SE, et al. Risk factors for dog
bites to owners in a general veterinary caseload. Appl Anim Behav Sci
2001;74:29–42.
17. Guy NC, Luescher UA, Dohoo SE, et al. A case series of bit-
ing dogs: characteristics of the dogs, their behaviour, and their vic-
tims. Appl Anim Behav Sci 2001;74:43–57.
18. Neidhart L, Boyd R. Companion animal adoption study.
J Appl Anim Welf Sci 2002;5:175–192.
19. Wells DL, Hepper PG. Prevalence of behaviour problems
reported by owners of dogs purchased from an animal rescue shelter.
Appl Anim Behav Sci 2000;69:55–65.
20. DiGiacomo N, Arluke A, Patronek G. Surrendering pets
to shelters: the relinquisher’s perspective. Anthrozoös 1998;11:
41–51.
21. New JC, Salman MD, Scarlett JM, et al. Moving: character-
istics of dogs and cats and those relinquishing them to 12 U.S. ani-
mal shelters. J Appl Anim Welf Sci 1999;2:83–96.
22. Scarlett JM, Salman MD, New JG, et al. The role of veteri-
nary practitioners in reducing dog and cat relinquishments and
euthanasias. J Am Vet Med Assoc 2002;220:306–311.
23. Flannigan G, Dodman NH. Risk factors and behaviors asso-
ciated with separation anxiety in dogs. J Am Vet Med Assoc 2001;
219:460–466.
24. Takeuchi YT, Houpt KA, Scarlett JM. Evaluation of treatments
for separation anxiety in dogs. J Am Vet Med Assoc 2000;217:342–345.
JAVMA, Vol 227, No. 11, December 1, 2005 Scientific Reports: Original Study 1761
SMALL ANIMALS
05-01-0034.qxp 11/16/2005 1:54 PM Page 1761
