Labradors in Backyards
Amanda Jane Kobelta, b, , , Paul H. Hemswortha, b, c, John L. Barnetta, c, Grahame J.
Colemana, d and Kym L. Butlerc
aAnimal Welfare Centre, 600 Sneydes Road, Werribee, Victoria 3030, Australia
bUniversity of Melbourne, 600 Sneydes Road, Werribee, Victoria 3030, Australia
cDepartment of Primary Industries, 600 Sneydes Road, Werribee, Victoria 3030,
Australia
dMonash University, Department of Psychology, PO Box 197, Caulfield East, Victoria
3145, Australia
Accepted 25 July 2006. Available online 6 September 2006.
Abstract
Most pet dogs in urban Australia are confined to their owners’ property in a suburban
backyard. Despite the literature on the effect of captive environments on both farm and
zoo animals, there is little objective information on the effects of confining pet dogs. The
prevalence of behavioural problems in pet dogs may be an indication that dogs are not
well adapted to some backyard environments. The aim of this study was to examine the
effect of being confined in backyards on dog behaviour by observing the behaviour of 55
Labrador retrievers in various backyard settings. The behaviour of the dogs was recorded
for 48 h and the social and physical environment of the dogs quantified. Time budgets of
dog behaviour were developed and relationships between behaviour and environmental
factors examined. Factors related to dog activity included the amount of foliage in the
yard, the number of transitions the dog made between locations and whether the dog was
kept inside at night, but also had a kennel provided in the yard. Factors related to problem
behaviours, such as digging and chewing, included whether the dog was gold coloured and
had no formal training, the dog being more active and the dog performing a high number
of transitions between locations in the backyard. It is suggested that it may be the type of
relationship with the owner that affects Labrador behaviour, rather than factors such as
size of the yard, having another dog present or time spent with the dog.
Keywords: Dogs; Behaviour problems; Suburban backyards; Environmental conditions
Article Outline
1. Introduction
2. Materials and methods
2.1. Choice of study dogs
2.2. Recruitment of participants
2.3. Eligibility of participants
2.4. Backyard assessment
2.5. Diary for dog owners
2.6. Video recording
2.7. Statistical analysis
3. Results
3.1. Household and dog demographics
3.2. Time budgets for dogs in backyards
3.3. Relationships between environmental factors and dog behaviours
4. Discussion
5. Conclusions
Acknowledgements
References
1. Introduction
Legislation in Victoria (Australia) requires that dogs be confined to the owners’ property
and not be allowed to wander at large (Domestic (Feral and Nuisance) Animals Act, 1994)
with similar requirements in most other states of Australia. In Australia the majority of
dogs housed in suburbia are confined to household backyards. It is well established that a
number of what may be considered as ‘behavioural abnormalities’ can occur in both farm
and zoo animals in confined environments particularly if the animal is isolated (Broom and
Johnson, 1993). Abnormal behaviours are defined as behaviours that are different in
either type or frequency from what is considered normal behaviour for that animal. Many
cases of abnormal behaviour in dogs have also been shown to occur under severe social
and environmental restriction (Lindsay, 2001). Dogs in backyards conditions are unlikely
to be severely restricted, but any tendency by dogs to show abnormal behaviour may
indicate either insufficient or excessive stimulation in their environment. Problem
behaviour can be normal or abnormal for the animal, but humans may consider it a
problem. Abnormal behaviours are instinctive or learnt (Beaver, 1999) and are often
suggested to be a result of dysfunctional learning (Lindsay, 2000). Behavioural problems
have become a major issue in pet dogs and can result in the dog being abandoned,
surrendered to shelters or euthanased ([Katcher and Beck, 1983] and [O’Farrell, 1990]).
There are many publications describing problem behaviour and its treatment, but little is
known of the aetiology of these behaviours. Furthermore, little is known of dog behaviour
in the backyard, particularly when confined for long periods of time. Often there is
physical evidence of the occurrence of behaviour, such as digging or chewing, but little is
known of the frequency or duration of these behaviours. Is the occurrence of such
abnormal behaviour a result of environmental factors in the backyard? Time budgets of
behaviour have been developed for dogs housed in shelters under various conditions (e.g.
[Hubrecht et al., 1992] and [Hetts et al., 1992]). There are also several studies on the
movements of feral free-ranging dogs (e.g. [Macdonald and Carr, 1995] and [Boitani et al.,
1995]). However, there are no studies on dogs in the backyard setting, which differs
markedly from these other situations.
The aims of this study were to identify and quantify the behaviour displayed by one breed
of dog confined to backyards and examine the relationships between these behaviours and
the conditions that make up the suburban backyard.
2. Materials and methods
2.1. Choice of study dogs
Studying one breed of dog reduces one of the possible variables that may affect dog
behaviour in the backyard, that is, breed differences. Labrador retrievers were chosen as
this pure breed of dog could be found in sufficient numbers in suburban Melbourne
(Australia).
2.2. Recruitment of participants
The Labrador Retriever Club of Victoria provided a list of breeders known to impose
similar early rearing practices on their puppies. Dog owners who obtained their dogs from
one of these breeders were contacted to inform them about the project and invite them to
participate when their dogs were about 18 months of age. If the owner was eligible (see
below for criteria) and agreed to participate, a time was made to meet them, assess the
backyard, provide a diary and instructions to complete it and for installation of the video
cameras. Recruitment was carried out over 2 years from February 2001 to January 2003.
2.3. Eligibility of participants
To be eligible to participate in the project, dog owners had to fulfil the following criteria;
have an average suburban backyard (nominated as less than 1/4 acre); have a Labrador
retriever between 17 and 19 months old; the dog must spend most of the day in the
backyard.
Households with either one or two Labradors or a Labrador with another breed of dog
were included in the study.
2.4. Backyard assessment
The participants were asked several questions about the dog's background, routine,
environment and whether they perceived their dog to have any problem behaviours. An
assessment of the backyard was then done with respect to size and the presence of specific
physical characteristics in the yard. The assessment included area of the yard; surface
type (grass, paved or combination), foliage (% of the surface area with plants other than
grass), shelter available, kennel present, bed present, toys present, other pets and whether
the dog was able to see into the house and/or into the street. The participant was also
given a diary to complete over 3 weeks (see below).
2.5. Diary for dog owners
The diary that dog owners were asked to complete consisted of six questions, repeated
daily. These questions were:
1. How much time was your dog in the backyard today?
2. How much time did your dog spend in the house today?
3. How much time today was your dog with you or another family member in the room of
the house or yard (minutes or hours)?
4. How much of the time, if any, was spent actively playing (e.g. throwing a ball in yard or
house)?
5. How much of the time, if any, was spent going for a walk/run?
6. How much time did your dog spend at home alone today (i.e. with no family member at
home)?
Records were requested to be in hours and minutes. Participants kept the diary for a 3-
week period, including the time that their backyard was videoed.
2.6. Video recording
Four black and white charge-coupled device (CCD) video cameras with audio were
installed in the backyard and connected to a four-channel video processor (quad-splitter)
and time-lapse video recorder. The backyard was videoed for 48 h at 24-h time lapse
speed (3 frames/s). If the dog was kept in the yard at night, red lights were used to provide
enough light for night recording. It has been suggested that red light does not affect dog
behaviour (Adams and Johnson, 1993). The tapes were scored according to location in the
backyard and the behaviour of the dog. The details of these elements are given in Table 1.
The behaviour and location of the dog in the yard were then transcribed continuously for
the 48 h using the Observer Video Pro Software (Noldus).
Table 1.
Descriptions of the locations and behaviours scored from video tapes Location
Dog out of yard
Dog in yard, not at door or gate
Dog at door of house
Dog at gate
Dog on its bed
Dog in its kennel
Dog at window of house
Dog at second/minor door of house
Behaviour
Dog not visible
Cannot see enough detail to define behaviour
Dog lying or sitting
Dog walking or running
Dog romping or playing
Dog walking around yard with nose to ground
Dog chewing object not toy, ball or food
Dog playing with/manipulating toy or ball
Dog consuming food
Dog consuming water
Dog urinating or defecating
Dog digging with front paws
Dog interacting with human
Dog interacting with another dog
Dog interacting with cat
Dog barking visual or aural
Dog licking, chewing or scratching self
Dog carrying or manipulating object not food
Dog standing on four or two legs
Dog chewing on bone
Dog carrying object or bone
Dog jumps up on two or four legs
2.7. Statistical analysis
Analyses between groups of dogs were carried out with Mann–Whitney U-tests.
Relationships between individual behaviours and variables in the backyards were
examined using Pearson's correlation analysis.
The percentage of time spent active was related to other measurements and factors
(excluding observed behaviours measurements) using multiple linear regression, after the
percentage of time had been angularly transformed. Interpretation of the multiple
regression was based on the most parsimonious model based on the measurements and
factors. Marginal adjustment, on the angular scale, was used in all tables of adjusted means.
The adjusted percentage of time spent active of individual points was calculated as the
backtransformation of the sum of the overall mean of the angularly transformed values
and the residual from the fitted model including only the other terms in the overall model.
The percentage of time showing problem behaviour was related to other measurements
and factors using multiple linear regression, after the percentage of time had been
angularly transformed.
The adjusted percentage of time in problem behaviour of individual points was calculated
as the backtransformation of the sum of the overall mean of the angularly transformed
values and the residual from the fitted model including only the other terms in the overall
model.
3. Results
3.1. Household and dog demographics
One hundred and twenty dog owners were contacted of which 57 (47.5%) agreed to be
involved in the study. This resulted in dog behaviour being studied in 55 yards. Of these,
51 households returned the dog owner diary.
Demographic information on the households was summarised as follows. Most households
(78%) had two adults and 45% had children. Only 16% of people were first time dog
owners (had never had a pet dog before) and all of these had only one dog. Most
households (80%) had one dog and most dogs were kept out at night (69%). The ratio of
male to female dogs was 1–1.1. Most dogs were desexed (78%). There were 28 black dogs,
20 gold and 8 chocolate coloured dogs.
The backyards ranged in size from 21 to 559 m2 and varied from small yards with a
concrete surface and no foliage (shrubs or trees in pots or in the ground) to quite complex
large yards with foliage, sheds and swimming pools. Backyard physical characteristics
were assessed as being present or not. Complexity was estimated by scoring (0, absent and
1, present) the following: foliage (>20% of yard surface covered), opportunity to see into
house, opportunity to see into street, presence of toys and presence of other pets,
resulting in a total score from 0 to 5. Shelter was not included in the analysis as all dogs
had some form of shelter.
Fifty-five percent of dogs had formal training, i.e. had attended obedience classes. Fifty-
five percent of gold, 50% of chocolate and 64% of black coloured dogs were trained. More
females (62%) than males (50%) were trained.
Sixty-five percent of dog owners stated that their dog showed some type of problem
behaviour. The main problem behaviours reported by owners were jumping up on people
(16%), digging (13%), pulling washed items off the clothesline in the backyard and chewing
them (11%) and chewing up plants (9%). Other behaviours that were reported at a lower
incidence included being disobedient (5%), aggression to people (4%) and aggression to
dogs (2%).
3.2. Time budgets for dogs in backyards
The activity of the dogs in the backyards over 24 h was plotted using the percentage of
each hour period that the dog was active (Fig. 1). There appeared to be two main peaks of
activity at around 0700–0900 h and again at 1700–1900 h.
Display Full Size version of this image (12K)
Fig. 1. Activity of dogs in the backyard over a 24-h period (N = 55).
To compare dog behaviour across yards, data were calculated for 10 consecutive hours in
each yard. This 10-h period commenced when the dog was first visible or active in the
morning. This allowed all dogs to be included in the analysis of activity from the point
when they were first visible or active in the morning and these results are presented in Fig.
2. Dogs spent the greatest proportion of the 10-h period inactive (74%) and out in the yard
(43%). The frequencies at which each behaviour and location were recorded was also
determined (Table 2).
Display Full Size version of this image (11K)
Fig. 2. Mean percentage of a 10-h period that dogs spent (a) in each behaviour and (b) in
each location (N = 55).
Table 2.
Mean frequency of occurrence (i.e. times recorded) in a 10-h period of each behaviour
and in each location (N = 55)
--------------------------------------------------------------------------------
Frequency
--------------------------------------------------------------------------------
Behaviour
Walking or running 140
Standing 83
Inactive (lying) 70
Exploring 40
Human interaction 14
Nonspecific 6
Carrying an object 5
Locations
In yard 82
At door 43
At gate 15
In kennel 11
On bed 10
At second door 5
Observations indicated that dogs were often most active in the presence of a human. Even
in yards that had two dogs, the dogs would start playing and interacting with each other
when the owner came out into the yard. When another dog or human was not present dogs
sometimes used an object or toy as a play partner. They play bowed at it and ran away
from it around the yard.
The number of transitions dogs made between locations in the 10-h period was examined
(Table 3). There was substantial variation between dogs in the number of transitions dogs
made between the yard and another location such as a door, gate or window. However,
most transitions were between the yard and the door of the house. The duration of these
transitions back to the same location were also calculated and it was found that 9% of
these transitions from the door back to the door were less than 10 s in duration. In fact in
one dog six transitions were short transitions involving the dog only circling around
before returning to the door.
Table 3.
The number of transitions dogs made between two locations in a 10-h period (N = 55)
Locations
--------------------------------------------------------------------------------
Mean
--------------------------------------------------------------------------------
Median
--------------------------------------------------------------------------------
Range
--------------------------------------------------------------------------------
Yard–door 23 19 1–82
Yard–gate 10 7 0–68
Yard–window 5 1 0–82
Yard–door 2 3 0 0–55
Total 41 33 1–145
From the diary data recorded by dog owners, the time that dogs spent in the house, in the
yard, with a person and out for walks was determined. The dogs spent from 3.5 to 24 h a
day in the backyard and correspondingly 0–20.5 h inside the house. Time spent with the
owner varied from 24 min to 20 h a day with dogs spending from 44 min to 9 h at home
alone. More than half the dogs (52%) were walked for half to 1 h a day. Seventeen dogs
(30%) were walked less than that or not at all.
The behaviour of dogs kept outside at night was examined over a 7-h period from 2200 to
0500 h. Thirty-eight dogs were kept outside at night, but two were not visible most of the
time (due to the red lights not operating), so data from 36 dogs were analysed. These dogs
spent most of the time inactive (96%) and in a kennel (65%). Ten percent of their time was
spent at the door of the house and 6% out in the yard. Dog behaviour did not appear to be
affected by the red lights in that activity did not increase, they did not look at the lights
nor were they deterred form entering their kennels with the lights shining on them.
3.3. Relationships between environmental factors and dog behaviours
A negative correlation was found between first-time dog ownership and whether the dog
obeyed commands (r = −0.39, P < 0.01). Dogs belonging to first-time dog owners did not
usually obey commands or were perceived not to do so by their owners. The more
children in the house, the less likely the dog was to have toys provided as shown by the
negative correlation between the dog having toys and the number of children (r = −0.51, P
< 0.01).
A positive correlation was found between whether the dog could see into the house and
whether it obeyed commands (r = 0.32, P < 0.05) suggesting that dogs that can see into
the house were more likely to obey commands. The time spent with a person per day was
positively correlated with time the dog spent in the house (r = 0.81, P < 0.01) and time the
owner spent playing with the dog (r = 0.49, P < 0.01) and negatively correlated with time
spent in the yard (r = −0.82, P < 0.01). These data suggest that the dogs spend most of the
time in the yard alone.
There was a positive correlation between the area of the yard and walk/run behaviour (r =
0.32, P < 0.05), carrying objects (r = 0.35, P < 0.01) and human interaction (r = 0.31, P <
0.05). This suggests that the larger the yard, the more dogs walk, run and carry objects
and there is also more human interaction.
To account for a possible “Monday” effect on dogs’ behaviour due to more human contact
on the weekend, dogs videoed on a Monday and Tuesday were identified. For these 19
dogs, the data from the first day of filming (Monday) were compared with the second day
(Tuesday). A Mann–Whitney U-test showed that the only significant difference in the two
groups was in chewing a bone. No other day effect on behaviour was detected.
The total number of transitions (r = 0.36, P < 0.01) and the number of transitions between
the yard and door (r = 0.37, P < 0.01) were also found to have positive correlations with
total activity.
The behaviours studied in the backyard were unable to be assigned to components via
principal component analysis as there were no outstanding factors and the components
were nonsensical. Further analysis was therefore carried out on the group ACTIVE
(calculated as 1 − proportion of time inactive). Behaviours considered to be possible
nuisance behaviours were grouped together for analysis. This group, called PROBLEM,
included barking, object manipulation, chewing objects, carrying objects and digging
behaviours. The mean percentage time dogs spent ACTIVE was 24%. Only two dogs
showed no problem behaviour and 30 dogs (55%) showed PROBLEM behaviours ranging
from 1 to 6% of the 10-h assessment period.
Due to the nonparametric nature of the data, Mann–Whitney U-tests were used to
ascertain whether there were any differences in factors such as dogs kept in or out at night
and yards with one or two dogs. The data used were proportion of visible time spent on
each behaviour over the 10-h period. There was a significant difference between yards
with one or two dogs in the time spent on ACTIVE behaviours (P = 0.015). Dogs were
more active if another dog was present. A comparison of dogs kept inside or outside at
night also showed a significant difference in time spent on ACTIVE (P = 0.035)
behaviours. Dogs that were kept inside at night were more active when in the yard.
A positive correlation was found between ACTIVE behaviour and the number of
transitions between locations (r = 0.36, P < 0.01), but not between PROBLEM behaviour
and number of transitions (r = 0.12, P > 0.05).
Total ACTIVE time of the dogs was found to be positively correlated with the presence of
more than 1% foliage in the yard (r = 0.33, P < 0.05), if the dog was both kept inside at
night and had a kennel in the yard (r = 0.39, P < 0.01) and whether there were one or two
dogs in the yard (r = 0.36, P < 0.01). A positive correlation was also found between the
ACTIVE and PROBLEM group of behaviours (r = 0.51, P < 0.01). A negative correlation
was found between the amount of time the dog was left alone and ACTIVE behaviours (r =
−0.31, P < 0.05).
These relationships were further analysed by linear regression and were found to be
significant in predicting the proportion of (angularly transformed) ACTIVE behaviour
(Fig. 3). The significant traits were: (1) the presence of more than 1% foliage in the yard (P
< 0.001), (2) whether the dog was inside at night and had a kennel in the yard (P < 0.001),
(3) whether the dog obeyed commands (P < 0.01) and (4) the total number of transitions
the dog made between locations in the yard (P < 0.001). This model accounts for 56.4% of
the variance in ACTIVE behaviour. Other backyard factors or behaviours had no
significant effect on total activity (ACTIVE) of the dogs.
Display Full Size version of this image (22K)
Fig. 3. Relationship between ACTIVE and the number of transitions between locations (N
= 55).
The PROBLEM group of behaviours was also subjected to an angular transformation
before analysis. There were positive correlations found between the proportion of time
showing PROBLEM behaviour and whether the dogs were gold in colour (r = 0.30, P <
0.01), and whether there was more than 1% foliage cover in the yard (r = 0.27, P < 0.05).
That is, dogs that were gold in colour and/or had more than 1% foliage cover in the yard
spent a significantly (P = 0.033) greater proportion of time displaying PROBLEM
behaviours.
These relationships were further analysed by linear regression and were found to be
significant in predicting the amount of PROBLEM behaviour. The significant traits were:
(1) the total percentage time ACTIVE (minus PROBLEM) (P < 0.001), (2) the dog was gold
in colour and had no training (P < 0.001) and (3) total number of transitions between
locations when the owner thought there was no PROBLEM behaviour (P < 0.01). These
relationships are demonstrated in Fig. 4, Fig. 5 and Fig. 6. This model accounts for 45.9%
of the variance in PROBLEM behaviour. Other backyard factors or behaviours had no
significant affect on PROBLEM behaviour of the dogs.
Display Full Size version of this image (21K)
Fig. 4. The percentage time dogs spend on problem behaviour (PROB) with respect to
colour and training (untransformed data) (N = 53).
Display Full Size version of this image (27K)
Fig. 5. The relationship between the proportion of time in problem behaviour and time in
total activity.
Display Full Size version of this image (26K)
Fig. 6. The relationship between the proportion of time in problem behaviour and the total
number of transitions between locations.
4. Discussion
This study identified and quantified, for the first time, the behaviour of Labrador
retrievers housed in suburban backyards. It also showed that there are relationships
between dog behaviour and some of the factors that are part of the backyard
environment. It is to be noted that comments made about dog behaviour and the factors
affecting it can only be specifically related to Labradors, but the considerable variation in
behaviour of these dogs suggests that some of these may be generalised to other dog
breeds. This study was also conducted on a relatively normal population of dogs. The
occurrence of actual problem behaviours was small and therefore what can be said about
the aetiology of problem behaviour is limited. Nevertheless, this study provides seminal
information on what dogs do while confined to the backyard of the house.
As expected there was a large variation in the type of backyard environment studied. This
was not just in physical characteristics, but in routine and in human and social contact that
the dog received. The yards varied in size, complexity and items provided for the dog. For
example, some dogs never went in the house or for a walk while others spent a lot of time
in the house and were walked daily. This variation was expected across suburban
Melbourne since a previous survey of dog owners (Kobelt et al., 2003) found a similar
result.
The behaviours indicated by dog owners to be a problem were jumping up on people,
digging and chewing, although the frequency was low. These results vary from those found
in surveys where aggressive behaviours were the problem behaviours most mentioned by
dog owners ([Voith et al., 1992] and [Beaver, 1994]). It has been suggested, however, that
Labrador retrievers are more prone to separation related behaviour problems (Bradshaw
et al., 2002) rather than aggression.
The dogs did not seem to be affected by the presence of the cameras or the red lights. Only
a couple of dogs responded to the cameras with one jumping up and barking at them. This
particular dog was not in the yard when the cameras were installed so may have been
responding to the novelty. While a couple of dog owners thought that their dogs were
affected by the red lights being on at night, video evidence did not support this. For
example, there was little orientation towards the lights, even early in the evening.
The behaviours of the dogs performed over a 48-h period were quantified. They spent
from half to nearly all their time inactive (45–96%) which is similar to the level of
inactivity that has been found to occur in shelter-housed dogs. Hubrecht et al. (1992)
found that all dogs in a shelter spent the majority of their time inactive (54–84%)
regardless of the spatial and social housing conditions. In the present study, the cause of
the observed peaks of activity cannot be determined as it may be due to circadian rhythms
or human activity, the latter could be that of the owners or of others outside of the
property. Berman and Dunbar (1983) found that free-ranging stray dogs travelled most
between 0600 and 1200 h. This suggests that the first peak of activity (0700–0900 h)
observed in the present study may be innate, but the second peak (1700–1900 h) may be
associated with human activity. Most of the dogs’ active time in the backyard was spent
walking, standing, exploring or chewing a bone if it was available.
From observing the tapes of the dogs it appeared that they were very focused on the
people present in the house. They appeared to spend a lot of time travelling between one
door of the house and another door, window or gate. This apparent focus on people was
not supported by any relationships between dog behaviour and time spent with people as
recorded in the diaries. However, there was a tendency for activity in the yard to increase
when the dog spent more time in the house.
Having more than one dog present in the yard or pen was found by Hubrecht et al. (1992)
and Hughes et al. (1989) to increase dog activity. Although in the present study the effect
was minimal after accounting for the effects of the variables in the regression model,
increased activity in pairs of dogs appeared to be due to the activity of one of the dogs.
That is, if one dog moved to investigate something the other dog would also move.
The area of the yard affected the amount of time dogs spent walking or running although it
did not affect total activity. The size of the enclosure has also been found to have no effect
on the total activity of captive wolves (Kreeger et al., 1996). The fact that area of the yard
was not directly related to dog activity is not surprising on the basis of viewing the
videotapes. Dogs spent the majority of their inactive time near the house, particularly
near the door or gate. The active time was often spent travelling between two doors or a
door of the house and a gate. That is, active dogs were changing locations frequently. This
fact accounts for the relationships found between area and walking or running as the dogs
in a large area often had further to travel between doors or the door and gate.
The amount of time dogs spent walking or running around could be affected by the
relationship between area and foliage where larger yards usually had more foliage. This
means that dogs are walking or running through the foliage not necessarily exploring it,
although the amount of foliage was related to the time spent exploring. The complexity of
the yard was also related to both foliage and barking behaviour. Although not examined in
this study, it is possible that yards that are more complex and have more foliage attract
other animals such as birds and dogs chasing and barking at birds are likely to be more
active. This example highlights the difficulties of isolating one factor as having a direct
effect on dog behaviour. The total activity in dogs was also increased by the presence of
foliage, which suggests that the presence of foliage may be a valuable source of
enrichment. If more than 1% of the yard had foliage, the dogs would spend time exploring
through it, digging in the soil and rubbing themselves against the foliage.
The relationships found in this study suggest that items previously suggested as
enrichment items, such as toys, did not affect the dogs’ behaviour, especially as the most
inactive dogs all had toys. This is contrary to the finding in shelter dogs by Hubrecht
(1993) where inactive time was reduced by “Gumabone” chew toys. However, this
behaviour may be more related to chewing a bone. Enrichment was unlikely to be as
important for dogs in this study since their environment was fairly diverse compared to
that of shelter or laboratory housed dogs. This was also suggested by Hetts et al. (1992)
where dogs in larger outdoor enclosures were less interested in available objects than dogs
in smaller or indoor enclosures. This would suggest that in general there is more
stimulation in larger, outdoor enclosures. In the present study, items such as food balls
that release food as the ball is manipulated were focused on as soon as they were given to
the dogs, but once they were empty (which only took minutes) no further interest was
taken in them. Having a bone to chew seemed to occupy the dogs for much longer (for
hours) and they would often return to it if was not completely eaten in the first bout.
It would appear from observing the tapes that dogs with the opportunity to observe the
street received more stimulation than dogs that did not, as they often responded to people
or dogs passing by. However, being able to see the street was not correlated with any of
the behaviours including barking.
It is feasible that dogs kept inside at night were more active during the day due to being
confined at night. This factor on its own, however, was only related to total activity of the
dogs when combined with having a kennel outside. There is no obvious explanation for the
relationship with the kennel being present. It is possible that dog owners who keep their
dog inside at night and also provide them with a kennel are a certain type of owner. For
example, they may be more attached to their dogs and consequently their dogs may be
more attached to them. These dogs also spent more time chewing bones presumably
because the owners provided them with bones or provided them more often than other
owners. Thus, these dogs may be more affected by being put out in the yard away from the
owner and this is reflected in the more active behaviour and more transitions between
locations. Their increased activity and transition frequency would indicate that these dogs
are spending a lot of time travelling between the door and a window or second door in an
attempt to keep visual contact with the people. If they cannot make contact and/or the
people go out, these dogs may redirect their behaviour to problem behaviours out of
frustration or isolation fear. This is suggested by the positive relationship found between
activity and problem behaviour and the relationship between these behaviours and the
combination of being kept inside at night plus having a kennel. This relationship, if it is
real, is indirect as there was no relationship found between problem behaviour and
transitions as no dogs showed a high number of transitions and a high percentage of time
showing problem behaviours. This effect of type of human interactions with dogs, as
opposed to the amount of time spent together, could also be suggested by the results
found by Podberscek and Serpell (1997) who found that low aggression dogs were
groomed by their owner more often than high aggression dogs. However, the amount of
time people spent playing with their dog did not differ between the two groups.
In this present study, it was also observed that dogs were most active when people were
present in the backyard. Although this did not make up much of the total active time
(<1%), most play behaviour (even between dogs) occurred when a person was present in
the yard. Hughes et al. (1989) also noted that dog activity was highest in the presence of
humans. In the present study, most play behaviour was observed in the presence of
humans, but some was directed at the other dog in two-dog yards and in a few instances in
single dog yards play bowing and romping was directed at a toy or object. If one dog was in
the yard it would solicit play with the owner usually by presenting a toy while if two dogs
were present they would play with a toy between them when the owner was present.
The fact that there was a relationship between problem behaviour and the total number of
transitions between locations only when the owner thought their dogs showed no problem
behaviours, is likely to be a result of owner perception rather than dog behaviour. This
may be due to people whose dogs were observed to show these behaviours not considering
them to have a problem. The alternative is that the behaviour that these people
considered to be a problem in their dog was not one observed on the tapes such as jumping
up on people.
The colour of the dog was the factor most highly correlated with the amount of time the
dog spent on problem type behaviours, with gold dogs spending the most time on these
behaviours especially if they were untrained. It has been noted that genetic influences
have a strong effect on many behavioural responses ([Wright and Nesselrote, 1987] and
[Serpell, 1987]). Dogs that are gold in colour may be more genetically predisposed to
being strongly attached to their owner and thus more anxious and prone to chew things
when isolated than dogs that are black or chocolate coloured. In a study by Houpt and
Willis (2001), it was found that chocolate Labradors were less likely to be presented for
problem behaviour than the other colours. Yellow (gold) dogs, however, were over
represented for aggression compared to the other colours. Other examples of dog colour
associated with behaviour include red/golden English Cocker Spaniels being more likely
to show aggression than black ones (Podberscek and Serpell, 1996).
The fact that having no training was only likely to increase problem behaviours in dogs if
they were gold may say more about the temperament of gold dogs than about the effects
of training. Podberscek and Serpell (1997) found that the type of training had no
significant effect on whether dogs had high or low aggression. Voith et al. (1992) also
found no affect of obedience training on the likelihood of dogs showing problem
behaviours.
The incidence of dogs chewing objects such as shoes was low and relied perhaps on
opportunity as well as tendency to do so. Examples of possible stereotypic behaviour also
occurred, but only in very few dogs. One dog spent a large proportion of its time walking
around the perimeter of the property and digging and chewing objects while two others
walked circular and “figure eight” patterns in front of the door to the house. Also this
study examined behaviours that are perceived to be a problem to owners such as digging
and chewing objects while other behaviour, such as greater inactivity, that occur a high
percentage of the time (90–96%) may be a sign of dysfunction/maladaption in some dogs.
5. Conclusions
This study has defined the elements that make up the physical and social environment of
dogs in suburban backyards. It has quantified the behaviour of Labrador retrievers
confined to this environment. As shown both in this and other studies, relationships are
apparent between some of the factors that make up this environment and the amount of
time dogs spent both active and on problem behaviours. Variables that were related to
active behaviour included the number of transitions dogs made between locations,
whether the dog was inside at night and was provided with a kennel and whether the owner
thought the dog was obedient. Time spent on problem behaviour was related to how active
the dogs were, together with dog colour and training.
This study provides seminal information on how dogs behave while confined to the
backyard of the house. It identifies that, for Labrador retrievers, factors often thought to
be important in affecting dog behaviour, such as the size of the yard, the presence of
another dog and the availability of toys, were less related to behaviour. A similar study
done on other breeds for comparison would be valuable in determining the importance of
breed on these behavioural problems.
Acknowledgements
This study was funded and supported by the Australian Research Council, Petcare
Information and Advisory Service, the RSPCA (Victoria) and Animal Behaviour Clinics.
The University of Melbourne, Human Research Ethics Committee (HREC No. 000441)
and the Animal Experimentation Ethics Committee (AEEC No. 00133) approved this
study. We acknowledge the advice and input of Robert Holmes (Animal Behaviour
Clinics), Peter Penson, Andrew Campbell and Russell McMurray (Bureau of Animal
Welfare, Victoria), Tony Cooke, Denise Humphries and Kristen Steel (Petcare
Information and Advisory Service), Rolf Bielharz (University of Melbourne), Peter
Barber (RSPCA) and Judith Blackshaw (University of Queensland). The fieldwork
associated with this study would not have been possible without the assistance provided
by Elaine Leeson. We would like to thank Sylvia Power and the Labrador Retriever Club of
Victoria for supplying lists of breeders with registered puppies for me to contact as well as
breeders for providing the names of their clients. Our great appreciation is given to the
dog owners that allowed us into their homes to video their dogs and took the time to
complete the diary of their dogs’ routines.
References
Adams and Johnson, 1993 G. Adams and K.G. Johnson, Sleep-wake cycles and other night-
time behaviours of the domestic dog Canis familiaris, Appl. Anim. Behav. Sci. 36 (1993),
pp. 233–248. Abstract
Beaver, 1994 B.V. Beaver, Owner complaints about canine behavior, J. Am. Vet. Med.
Assoc. 204 (1994), pp. 1953–1955. View Record in Scopus | Cited By in Scopus (23)
Beaver, 1999 B.V. Beaver, Canine Behavior: A Guide for Veterinarians, W.B. Saunders
Co., Philadelphia (1999).
Berman and Dunbar, 1983 M. Berman and I. Dunbar, The social behaviour of free-ranging
suburban dogs, Appl. Anim. Ethol. 10 (1983), pp. 5–17.
Boitani et al., 1995 L. Boitani, F. Francisci, P. Ciucci and G. Andreoli, Population biology
and ecology of feral dogs in central Italy. In: J. Serpell, Editor, The Domestic Dog: Its
Evolution, Behaviour and Interaction with People, Cambridge University Press,
Cambridge (1995), pp. 217–244.
Bradshaw et al., 2002 J.W.S. Bradshaw, J.A. McPherson, R. Casey and I. Larter, Aetiology
of separation-related behaviour in domestic dogs, Vet. Rec. 151 (2002), pp. 43–46. View
Record in Scopus | Cited By in Scopus (8)
Broom and Johnson, 1993 D.M. Broom and K.G. Johnson, Stress and Animal Welfare,
Chapman and Hall, London (1993).
Hetts et al., 1992 S. Hetts, D.J. Clark, J.P. Calpin, C.E. Arnold and J.M. Mateo, Influence
of housing conditions on beagle behaviour, Appl. Anim. Behav. Sci. 34 (1992), pp. 137–
155. Abstract
Houpt and Willis, 2001 K.A. Houpt and M. Willis, Genetics of behaviour. In: A. Ruvinsky
and J. Sampson, Editors, The Genetics of the Dog, CABI Publishing, Wallingford (2001),
pp. 371–400.
Hubrecht, 1993 R.C. Hubrecht, A comparison of social and environmental enrichment
methods for laboratory housed dogs, Appl. Anim. Behav. Sci. 37 (1993), pp. 345–361.
Abstract
Hubrecht et al., 1992 R.C. Hubrecht, J.A. Serpell and T.B. Poole, Correlates of pen size
and housing conditions on the behaviour of kennelled dogs, Appl. Anim. Behav. Sci. 34
(1992), pp. 365–383. Abstract
Hughes et al., 1989 H.C. Hughes, S. Campbell and C. Kenney, The effects of cage size and
pair housing on exercise of beagle dogs, Lab. Anim. Sci. 39 (1989), pp. 302–305. View
Record in Scopus | Cited By in Scopus (21)
Katcher and Beck, 1983 A.H. Katcher and A.M. Beck, New Perspectives on Our Lives with
Companion Animals, University of Pennsylvania Press, Philadelphia (1983).
Kobelt et al., 2003 A.J. Kobelt, P.H. Hemsworth, J.L. Barnett and G.J. Coleman, A survey
of dog ownership in suburban Australia: conditions and behaviour problems, Appl. Anim.
Behav. Sci. 82 (2003), pp. 137–148. SummaryPlus | Full Text + Links | PDF (144 K) |
View Record in Scopus | Cited By in Scopus (15)
Kreeger et al., 1996 T. Kreeger, D. Pereira, M. Callahan and M. Beckel, Activity patterns
of gray wolves housed in small versus large enclosures, Zoo Biol. 15 (1996), pp. 395–401.
Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)
Lindsay, 2000 S.R. Lindsay, Handbook of Applied Dog Behavior and Training, vol. 1:
Adaptation and Learning, Iowa State University Press, Iowa (2000).
Lindsay, 2001 S.R. Lindsay, Handbook of Applied Dog Behavior and Training, vol. 2:
Etiology and Assessment of Behavior Problems, Iowa State University Press, Iowa (2001).
Macdonald and Carr, 1995 D. Macdonald and G. Carr, Variation in dog society: between
resource dispersion and social flux. In: J. Serpell, Editor, The Domestic Dog: Its
Evolution, Behaviour and Interactions with People, Cambridge University Press,
Cambridge (1995), pp. 199–216.
O’Farrell, 1990 V. O’Farrell, Behavioural problems in companion animals. In: P.
Monaghan and D. Woodgush, Editors, Managing the Behaviour of Animals, Chapman and
Hall, New York (1990), pp. 233–252.
Podberscek and Serpell, 1996 A.L. Podberscek and J.A. Serpell, The English Cocker
Spaniel: preliminary findings on aggressive behaviour, Appl. Anim. Behav. Sci. 47 (1996),
pp. 75–89. SummaryPlus | Full Text + Links | PDF (1038 K)
Podberscek and Serpell, 1997 A.L. Podberscek and J.A. Serpell, Environmental influences
on the expression of aggressive behaviour in English Cocker Spaniels, Appl. Anim. Behav.
Sci. 52 (1997), pp. 215–227. SummaryPlus | Full Text + Links | PDF (911 K) | View Record
in Scopus | Cited By in Scopus (26)
Serpell, 1987 J.A. Serpell, The influence of inheritance and environment on canine
behaviour: myth and fact, J. Small Anim. Pract. 28 (1987), pp. 949–956.
Voith et al., 1992 V.L. Voith, J.C. Wright and P. Danneman, Is there a relationship between
canine behavior problems and spoiling activities, anthropomorphism and obedience
training, Appl. Anim. Behav. Sci. 34 (1992), pp. 263–272. Abstract
Wright and Nesselrote, 1987 J.C. Wright and M.S. Nesselrote, Classification of behavior
problems in dogs: distribution of age, breed, sex and reproductive status, Appl. Anim.
Behav. Sci. 19 (1987), pp. 169–178. Abstract
Corresponding author at: 621 Sneydes Road, Werribee, Victoria 3030, Australia. Tel.:
+61 3 9742 8788; fax: +61 3 9742 8700.
Applied Animal Behaviour Science
Volume 106, Issues 1-3, August 2007, Pages 70-84
| The behaviour of Labrador retrievers in suburban backyards: The relationships between the backyard environment and dog behaviour |