Dog training methods: Their use, effectiveness and interaction with behaviour and welfare

Abstract

Historically, pet dogs were trained using mainly negative reinforcement or punishment,
but positive reinforcement using rewards has recently become more popular. The
methods used may have different impacts on the dogs’ welfare. We distributed a
questionnaire to 364 dog owners in order to examine the relative effectiveness of
different training methods and their effects upon a pet dog’s behaviour. When asked how
they trained their dog on seven basic tasks, 66% reported using vocal punishment, 12%
used physical punishment, 60% praise (social reward), 51% food rewards and 11% play.
The owner’s ratings for their dog’s obedience during eight tasks correlated positively
with the number of tasks which they trained using rewards (P < 0.01), but not using
punishment (P = 0.05). When asked whether their dog exhibited any of 16 common
problematic behaviours, the number of problems reported by the owners correlated with
the number of tasks for which their dog was trained using punishment (P < 0.001), but
not using rewards (P = 0.17). Exhibition of problematic behaviours may be indicative of
compromised welfare, because such behaviours can be caused by—or result in—a state
of anxiety and may lead to a dog being relinquished or abandoned. Because punishment
was associated with an increased incidence of problematic behaviours, we conclude that
it may represent a welfare concern without concurrent benefits in obedience. We suggest
that positive training methods may be more useful to the pet-owning community

Full text

63
© 2004 Universities Federation for Animal Welfare
The Old School, Brewhouse Hill, Wheathampstead,
Hertfordshire AL4 8AN, UK
Animal Welfare 2004, 13: 63-69
ISSN 0962-7286
Dog training methods: their use, effectiveness and interaction with
behaviour and welfare
EF Hiby*, NJ Rooney and JWS Bradshaw
Anthrozoology Institute, Department of Clinical Veterinary Science, University of Bristol, Langford, Bristol BS40 5DT, UK
* Correspondence: elly.hiby@bristol.ac.uk
Abstract
Historically, pet dogs were trained using mainly negative reinforcement or punishment, but positive reinforcement using rewards has
recently become more popular. The methods used may have different impacts on the dogs' welfare. We distributed a questionnaire
to 364 dog owners in order to examine the relative effectiveness of different training methods and their effects upon a pet dog's
behaviour. When asked how they trained their dog on seven basic tasks, 66% reported using vocal punishment, 12% used physical
punishment, 60% praise (social reward), 51% food rewards and 11% play. The owner's ratings for their dog's obedience during eight
tasks correlated positively with the number of tasks which they trained using rewards (P < 0.01), but not using punishment (P =0.5).
When asked whether their dog exhibited any of 16 common problematic behaviours, the number of problems reported by the
owners correlated with the number of tasks for which their dog was trained using punishment (P < 0.001), but not using rewards
(P = 0.17). Exhibition of problematic behaviours may be indicative of compromised welfare, because such behaviours can be caused
by  or result in  a state of anxiety and may lead to a dog being relinquished or abandoned. Because punishment was associated
with an increased incidence of problematic behaviours, we conclude that it may represent a welfare concern without concurrent
benefits in obedience. We suggest that positive training methods may be more useful to the pet-owning community.
Keywords:animal welfare, domestic dog, humananimal interaction, obedience, problematic behaviour, training methods
Introduction
Domestic dogs (Canis familiaris) are kept in large numbers
by humans worldwide, and their welfare is worth serious
ethical and scientific consideration. The majority of previous
research in this area has been in laboratories and re-homing
centres (eg Hubrecht 1993, 1995; Hennessey et al 1998;
Wells & Hepper 2000), but most domestic dogs are kept as
pets within homes. The manner in which these dogs are
trained may be of particular importance to their welfare.
Most pet dogs receive at least rudimentary training. For
instance, dogs are trained to eliminate in an appropriate
place, to walk to heel, and to obey a range of commands
including sitting and lying down. The methods by which
these tasks are trained vary greatly.
Traditional training techniques have used mainly aversive
stimuli, either in the form of positive punishment (applica-
tion of an aversive stimulus in response to an undesirable
behaviour) or negative reinforcement (removal of an aver-
sive stimulus leading to an increase in the performance of a
desirable behaviour) (Lieberman 1999). The use of aversive
stimuli in training may have negative welfare implications:
it is thought to cause suffering (Beerda et al 1997), possibly
poses health risks (through increased levels of physiological
stress), and has been found to be related to aggression
towards other dogs (Roll & Unshelm 1997).
In contrast, other training methods utilise positive reinforce-
ment through the presentation of rewards in response to
desired behaviours. These methods have been adopted as
part of the training regime of many working dog organisa-
tions, including the Guide Dogs for the Blind Association
(Johnston 1995) and agencies training dogs to detect illegal
substances (Adams & Johnson 1994).
In this paper we survey the training methods used by the
general pet owning community in the UK. We also examine
the relative effectiveness of these methods for training a
variety of specific tasks. This is important because positive
training methods are likely to be adopted only if they are at
least as successful as their negative counterparts. We
hypothesise that specific training methods will be associated
with the performance of behaviours deemed problematic by
owners (Overall 1997). The manifestation of problematic
behaviours may be important for dog welfare for several
reasons: first, they are often caused by, or result in, states of
anxiety; and second, problematic behaviours have been
implicated in the weakening of the petowner relationship
(Mugford 1981), which may result either in the relinquish-
ment of dogs (Serpell 1996) or in euthanasia.
Previous studies of the relationship between training methods
and problematic behaviours have yielded apparently con-
flicting results. Podberscek and Serpell (1997) and Voith et
al (1992) found no relationship between obedience training
and problematic behaviours, while Clark and Boyer (1993)
and Jagoe and Serpell (1996) found a decrease in these
behaviours following obedience training. This discrepancy
Universities Federation for Animal Welfare Science in the Service of Animal Welfare

64 Hiby et al
may be because these studies compared formal (under the
supervision of a dog trainer) versus informal training and
took no account of the specific methods used (eg punishment
and reward). The aim of the current study is to document the
use of training methods by the pet-owning community and
investigate how these methods interact with both obedience
and problematic behaviours.
Methods
Participants and distribution
Three main sites were used in Hampshire: Southampton
Common (50°55´36´´N, 1°24´39´´W), Southampton Sports
Centre (50°56´28´´N, 1°25´14´´W), and Deers Leap
(50°52´53´´N, 1°30´45´´W). Four main sites were used in
Cambridgeshire: Gog Magog Down (52°9´22´´N,
0°10´25´´E), Cherry Hinton Estate Park (52°11´13´´N,
0°9´59´´E), Midsummer Common (52°12´36´´N, 0°7´46´´E),
and Jesus Green (52°12´43´´N, 0°7´25´´E). These sites were
chosen for their high concentration of dog walkers and
because they represented a cross-section of the types of
areas open for dog walking, ie both rural and urban. All sites
were visited at various times of the day during daylight
hours, on both weekends and weekdays. All dog walkers
were approached by the experimenter and asked if they
would mind participating in a questionnaire survey about
their dog. If amenable, they were given a questionnaire with
a stamped addressed envelope.
In addition, 60 questionnaires were distributed to local vet-
erinary surgeries and pet stores in each of the two counties
(Hampshire and Cambridgeshire) for voluntary pick-up.
This ensured that not only people who walked their dogs
received the questionnaire.
To avoid reports of puppy behaviour, only people who
owned a dog over one year of age were recruited. If respon-
dents owned more than one dog, they were asked to answer
with regard to the youngest dog (over one year), as it was
assumed that the training techniques used for the younger
dog would be remembered better.
Design
The questionnaire contained 26 questions, 13 simple/multiple-
choice and 13 open-ended, covering the following topics.
Demographics
Respondents were asked their gender and age and that of
their dog. They also gave information on the breed, sexual
status (entire or neutered), source and age at acquisition of
the dog, and how many dogs they owned both presently and
in the past.
Training methods
Respondents were asked open-ended questions about the
training methods that they had used to train seven common
tasks. They were first asked about three specific training
situations: how they toilet trained their dog, how they react-
ed to their dog chewing household objects, and how they
reacted if their dog stole food or other objects. They were
then asked which methods they had used to train their dog
to perform four tasks: to come when called, to sit on com-
mand, to leave or give up an object on command, and to
walk to heel.
Obedience
Owners were asked to rate their dogs obedience on a scale
of 15 (5 being the most obedient) for each of the seven
tasks. They were also asked to give an eighth obedience
score (15) for overall obedience. All eight obedience
scores were summed to give a summed obedience score.
Problematic behaviours
Respondents were presented with a list of 16 common prob-
lematic behaviours (Barlow 2003) and asked to indicate
whether their dog had shown each of them in the past,
currently or never. These 16 problems were reduced to 13
during analysis (Table 1).
Statistical analysis
All statistical analyses were carried out using SPSS 11 for
Windows (SPSS Inc, 19892001).
Demographics
The data from Hampshire and Cambridgeshire were com-
pared using the χ2test for categorical data and the Mann-
Whitney Utest (Siegel & Castellan 1988) for ordinal data.
Training methods and obedience
Questionnaires with incomplete training sections, or in
which the respondents stated that they had not trained the
dog themselves, were excluded from analysis. We analysed
the success of each commonly used method for training the
seven tasks. For each task in turn, we examined each training
method mentioned by at least 10% of respondents. We used
Mann-Whitney Utests to compare the obedience scores
given by those respondents who had used that training
method with the obedience scores given by those respon-
dents who had not used it.
Each reported training method was then categorised as
reward-based, punishment-based, or miscellaneous (ie not
obviously rewarding nor punishing) (see Table 2). The
authors and also two clinical animal behaviourists agreed
the categorisation. For each respondent, we then added up
the total number of times that they had reported use of each
method type (reward, punishment and miscellaneous) to
generate three new variables: reward frequency, punishment
frequency, and miscellaneous method frequency. These
variables were tested for correlation with the summed obe-
dience scores using Spearmans Rank correlation tests
(Siegel & Castellan 1988).
Finally, we categorised the respondents according to their
general training strategy. Each respondent was categorised
as using reward only, punishment only, a combination of
reward and punishment, or miscellaneous methods only.
The summed obedience scores for the four groups were
compared using the Kruskal-Wallis test (Siegel & Castellan
1988).
© 2004 Universities Federation for Animal Welfare

Dog training methods 65
Problematic behaviours
All problematic behaviours reported by less than 10% of the
population were combined with other problem categories.
Nipping and growling at people were combined into the cat-
egory aggression towards people. Destruction, noise and
elimination (urinate or defecate) when left alone were com-
bined into the category separation-related behaviours.
This resulted in 13 problematic behaviour categories
(Table 1).
We tested the relationship between the frequency of reward,
punishment and miscellaneous methods and the number of
current problematic behaviours using Spearmans rank
correlation tests. We initially used the number of current
problematic behaviours, as those performed in the past
might have been influenced by a previous owners treat-
ment or reflect normal puppy behaviour, and might not be
related to the current owners training regime. Previous
problems were only included, to increase sample sizes,
when exploring specific relationships between individual
problems and training methods.
Respondents using each of the four training strategies
(reward only, punishment only, a combination of reward
and punishment, and miscellaneous methods only) were
compared for the total incidence of the 13 problems among
their dogs using Kruskal-Wallis tests. Next, χ2tests were
used to compare the number of dogs undergoing each train-
ing strategy with their performance of individual problem-
atic behaviours (currently, in the past, or never). Where non-
significant results were found, the dogs in those training
strategy groups were combined and the χ2test re-run. This
process was continued until only those training strategies
that showed a significantly different incidence of the prob-
lematic behaviour remained.
Results
Returned questionnaires
600 questionnaires were distributed (300 in each of the two
counties) and 364 were returned (return rate 61%). Of these,
38 were either incomplete or were answered for a dog under
one year old. Of the remaining 326, exactly 50% came from
each county.
Demographics
Of the respondents, 69.6% were female, 26.1% were male,
and 4.3% of questionnaires were answered by a mixed-sex
couple. Overall 4.6% of respondents were under 25 years
old, 26.1% were between 25 and 40, 46.9% were between
41 and 60 and 22.2% were over 60 years old.
The dogs ranged from 1 to 15 years with a mean age of 61
months (± 40 months); 22.4% were entire males, 32.5%
were neutered males, 7.7% were entire females and 37.4%
were neutered females. The mean number of dogs owned by
the respondents at the time of receiving the questionnaire
was 1.4 (± 0.8), and the mean number of dogs owned in the
past was 2.3 (± 3.7).
Breeds were split into Kennel Club categories (see
Discover Dogs at www.the-kennel-club.org.uk), with an
extra category containing all cross-breeds including
lurchers. In total, 1.8% of dogs were hounds, 4.3% were
working breeds, 17.5% were terriers, 31.9% were gun dogs,
14.7% were pastoral, 2.8% were utility, 5.8% were toy
breeds and 21.1% were cross-breeds. The most popular
breed was the Labrador Retriever (10.4%).
When the data from Hampshire and Cambridgeshire were
compared, two significant differences were found: a greater
proportion of Cambridgeshire respondents were women
(73% versus 66.3%; χ2= 21.9, P< 0.01); and
Cambridgeshire respondents had owned a greater number of
Animal Welfare 2004, 13: 63-69
Table 1 Incidence of the 13 categories of problematic behaviours, displayed as a percentage of the total respondent
population.
Problematic behaviour Never displayed Displayed in the past Currently being displayed
Barking at people 15.8 8.2 75.7
Aggression towards people 50.2 16.8 33.0
Barking at dogs 34.4 12.9 52.7
Nipping at dogs 74.8 7.7 13.8
Growling at dogs 40.1 17.5 42.4
Fear in a few situations 22.0 17.8 60.2
Fear in many situations 46.6 19.8 33.6
Excitement in a few situations 9.6 8.3 82.1
Excitement in many situations 37.1 15.1 47.8
Separation-related behaviours 46.3 38.0 15.6
Inappropriate mounting 71.3 14.7 14.1
Repetitive behaviours 80.1 10.1 9.8
Eating non-foodstuffs 57.6 25.9 16.5

66 Hiby et al
dogs previously (median 2 versus 1; U= 11676.5,
P= 0.05). Since no further differences were found, we con-
sidered the two populations to be sufficiently similar to
combine them for further analysis.
Training methods
Twelve training techniques were described (by >10% of the
respondents) in response to the seven training tasks. When
methods were categorised into the three types (Table 2), the
type of training method was seen to vary depending on the
task being trained. Punishment was commonly reported
when training dogs not to chew household objects or steal;
rewards were more common when training dogs to sit or
come to call (Table 3).
Overall 20.2% of respondents used reward-based methods
only, 9.8% used punishment-based methods only, 60.4%
used a combination of reward and punishment, and 9.6%
used miscellaneous methods only or mentioned no methods
at all.
Obedience
The median summed obedience score was 33 (first quar-
tile = 29, third quartile = 36). The scores ranged from 19 to
the maximum possible score of 40.
Obedience and overall use of training methods
The reward frequency correlated positively with the
summed obedience score (Rho = 0.26, P< 0.01). No signif-
icant correlation was found between obedience and the fre-
quency of either punishment or miscellaneous methods.
The summed obedience scores for the four categories of
respondents (those that used reward only, punishment only,
a combination of reward and punishment, or miscellaneous
methods only) differed significantly (Kruskal-Wallis;
χ2= 8.152, P< 0.05). Highest obedience scores were given
by respondents using reward-based methods only; this was
followed by those using a combination of reward and
punishment, and then those using punishment only.
Respondents using miscellaneous methods only reported
the lowest obedience scores.
Obedience and training for each task
When the obedience scores for each individual task were
compared between respondents who had and who had not
used each of the common methods, three significant results
were found:
(1) Obedience to leave or give up an object was significant-
ly greater for those dogs that had been trained using play as
a reward, in comparison with those dogs that had not been
trained using this method (U= 1977.5, P< 0.01).
(2) Obedience to walk to heel was significantly greater for
those dogs that had been trained using praise as a reward, in
comparison with those dogs that had not been trained using
this method (U= 1612, P< 0.05).
(3) Obedience for not chewing household objects was
greater in those dogs that had received an alternative object
to chew in response to their chewing behaviour, in compar-
ison with those dogs that had not received an alternative
(U= 3181.5, P< 0.05).
No significant associations were found between the obedi-
ence scores for any of the other tasks and any specific train-
ing technique.
Problematic behaviours
Of the respondents, 97.2% mentioned at least one of the 13
problematic behaviours. The most prevalent problems were
showing excitement in a few situations, barking at people,
and showing fear in a few situations (Table 1).
Problematic behaviours and use of training methods
The punishment frequency correlated positively with the
number of current problematic behaviours (Rho = 0.17,
P< 0.01). In contrast, no significant correlations were
© 2004 Universities Federation for Animal Welfare
Table 2 Categorisation of the 12 training techniques into the three training method types.
Training method category Training technique
Punishment-based Physical punishment (eg smacking, tapping nose)
Vocal punishment (eg shouting, using stern voice)
Sending the dog to bed/outside
Tugging back at lead in heel training
Reward-based Play reward
Food reward
Praise reward
Reward, type unspecified
Miscellaneous Newspaper on floor to encourage elimination in a particular area during toilet training
Providing an alternative object to chew
Ignoring the behaviour (chewing or stealing behaviour)
Placing the dog into sit using pressure on hindquarters

Dog training methods 67
found between the frequency of either reward or miscella-
neous methods and the incidence of problematic behaviours.
The number of current problematic behaviours reported by
the four categories of respondents (those that used reward
only, punishment only, a combination of reward and punish-
ment, or miscellaneous methods only) differed significantly
(Kruskal-Wallis; χ2= 8.993, P< 0.05). The greatest number
of current problematic behaviours was reported by respon-
dents using punishment only, or a combination of both pun-
ishment and reward (both medians = 5); the lowest number
was reported by respondents using reward only or miscella-
neous methods only (both medians = 4).
When the respondents using the four different training
strategies were compared for the occurrence of each of the
13 problematic behaviours, three significant associations
were found:
(1) Owners reporting reward-based methods only had the
lowest percentage of dogs showing current over-excitement
(χ2= 11.6, P< 0.01).
(2) Owners using punishment, either alone or in combina-
tion with reward, had the highest percentage of dogs
exhibiting separation-related problems either currently or in
the past (χ2= 29.2, P< 0.001).
(3) Owners who reported using only miscellaneous meth-
ods, or no methods at all, reported the lowest percentage of
dogs eating non-foodstuffs either currently or in the past
(χ2= 6.8, P< 0.01).
Discussion
This survey has shown that reward-based training is used
extensively within the dog-owning community, with over
three-quarters of respondents reporting using some form of
reward. However, it is unusual for owners to base all of their
training on rewards, and the vast majority use a training
regime that combines reward and punishment. Thus, pun-
ishment (verbal or physical) continues to feature prominent-
ly in the training of pet dogs.
Although there is an increasing concern that certain forms
of punishment can cause suffering (Beerda et al 1997),
there remains a general belief that, for many canine tasks,
punishment is the most effective training technique
(Christiansen et al 2001; Marschark & Baenninger 2002).
For example, many owners used punishment-based training
to teach dogs not to chew or steal objects. However, this
survey suggests that for everyday training, punishment is
not the most effective method. Furthermore, for certain tasks,
reward-based methods are significantly more successful.
When we compared dogs obedience at seven basic tasks,
for four of these tasks (toilet training, stopping stealing
objects, coming to call and sitting on command) we discov-
ered no significant difference between dogs trained using
one specific method in comparison with another. However,
for three tasks (leaving or giving up an object on command,
walking to heel and avoiding chewing household objects),
the use of specific methods was associated with significantly
higher obedience scores. Although the most effective tech-
nique varied according to the specific training task, for none
of the tasks was a punishment-based method most effective.
Even for chewing and stealing objects, where punishment is
very commonly used, those owners who used it did not
report greater obedience. Thus, examination of the individ-
ual tasks provides no support for the value of punishment.
Further evidence supporting the use of reward-based
methods was found when examining overall obedience
scores. These scores correlated significantly with the
number of times the owners reported using reward-based
training methods, but were unrelated to their reports of
either punishment-based or miscellaneous methods.
Furthermore, dogs trained exclusively using reward-based
methods were reported to be significantly more obedient
than those trained using either punishment or a combination
of reward and punishment.
These results suggest that there is a link between the use of
reward-based methods and obedience in pet dogs. However,
this is a correlation and although the frequent use of rewards
may lead to increased obedience, it is also possible that
when dogs show an initial high level of obedience, their
owner is more inclined to use reward-based training
methods. However, regardless of which is cause and which
is effect, it is clear that use by the general dog-owning com-
munity of punishment-based methods, as compared to reward-
based methods, does not result in a more obedient dog.
Animal Welfare 2004, 13: 63-69
Table 3 Percentage of respondents using each training method type for each of the seven tasks.
Training situation Training method categories
Punishment Reward Miscellaneous
Toilet training 11.6 39.1 44.9
Chewing household objects 78.5 4.3 39.8
Stealing food/objects 83.6 7.3 10.3
Sit on command 0.4 75.2 39.1
Come to call 1.9 77.8 0
Leave/give-up object 29.8 63.3 0
Heel training 26.2 45.2 0

68 Hiby et al
Obedience is an important attribute of a dogowner rela-
tionship. In previous studies, Clark and Boyer (1993) found
that obedience training, with its concurrent increase in task
obedience, was related to an improvement in the
humancanine relationship. Also, Serpell (1996) found
that owners were less attached to their newly adopted dog if
there was a large discrepancy between their ratings of the
ideal and the actual dog, and one of the main discrepancies
was lack of obedience. Because satisfied owners are less
likely to relinquish or abandon their dogs (Arkow & Dow
1984), training methods that produce an obedient dog may
exert a secondary welfare benefit.
Problematic behaviours are common within the general dog
population (Voith et al 1992; Clark & Boyer 1993; OFarrell
1997). In this survey, 97.2% of owners reported their dogs
showing behaviours which they themselves did not neces-
sarily regard as a problem, but others might. Our data sug-
gest that the use of specific training methods may be linked
to enhanced exhibition of problematic behaviours. The
number of times owners reported using punishment-based
methods correlated positively with the number of potentially
problematic behaviours they reported. Furthermore, those
owners who trained their dogs using a regime based entirely
on punishment, or a combination of punishment and reward,
reported significantly more problems than those using only
reward-based or miscellaneous methods. It may be that pun-
ishment increased the number of problematic behaviours
displayed, perhaps by creating a state of anxiety or conflict
in the dog that is later expressed as a problematic behaviour.
However, it is also possible that owners of dogs that already
exhibit many problematic behaviours are more likely to
incorporate punishment into their training regime. Either
way, at the time of completion of the questionnaire,
increased punishment correlated with increased problematic
behaviours, which suggests that punishment had not effec-
tively eliminated these behaviours.
Previous research has found a link between punishment-
based Schutzhund training and dogdog aggression (Roll &
Unshelm 1997). We did not detect this, but we did find a
link between the use of punishment and increased incidence
of separation-related behaviour. Separation-related prob-
lems are known to have many causes, including anxiety and
conflict (eg Askew 1996), which may be exacerbated by the
use of punishment.
In addition, we saw a reduced incidence of over-excitement
in dogs trained using reward-based methods only.
Punishment or combined methods can lead to anxiety, which
may be manifest as over-excitement; hence, a predictable
reward-based regime may be more effective at reducing
excitement. The relationship between the use of miscella-
neous methods and a decreased incidence of eating non-food-
stuffs does not appear to have a straightforward explanation.
Overall, our results suggest that punishment-based training
is not effective at reducing the incidence of problematic
behaviours, and its use seems to be linked with the
increased occurrence of potential problems. In contrast, the
frequency of rewards was unrelated to problematic behaviours.
Although our survey has found no support for the use of
punishment-based methods, we are not suggesting that they
cannot be used effectively. Historically, punishment has
been used successfully for many types of dog training (Most
2000). However, within the pet-owning community, train-
ing is very often performed by inexperienced people who
are unfamiliar with the behavioural principles involved,
hence the timing of the delivery of punishment may be inap-
propriate and its use inconsistent. We suggest that such
training can result in states of anxiety in the animal, leading
to an increased probability of problematic behaviours and
inhibiting the desired increase in obedience. Thus, for the
general dog-owning population, reward-based training
methods may produce a more balanced and obedient animal.
Conclusions and welfare benefits
There are ethical concerns that dog-training methods incor-
porating physical or verbal punishment may result in pain
and/or suffering. We provide evidence that, in the general
dog-owning population, dogs trained using punishment are
no more obedient than those trained by other means and,
furthermore, they exhibit increased numbers of potentially
problematic behaviours. Problematic behaviours can com-
promise welfare as they are often associated with an
increased state of anxiety (eg Askew 1996) and they can also
lead the owner to relinquish the dog (Serpell 1996). Because
reward-based methods are associated with higher levels of
obedience and fewer problematic behaviours, we suggest
that their use is a more effective and welfare-compatible
alternative to punishment for the average dog owner.
Acknowledgements
We would like to acknowledge the Defence Science and
Technology Laboratory for funding this project and the
many respondents who took the time to complete our survey.
References
Adams GJ and Johnson KG 1994 Sleep, work, and the effects
of shift work in drug detection dogs, Canis familiaris. Applied Animal
Behaviour Science 41: 115-126
Arkow PS and Dow S 1984 The ties that do not bind: a study
of the humananimal bonds that fail. In: Anderson RK, Hart BL
and Hart LA (eds) The Pet Connection: Its Influence on Our Health
and Quality of Life pp 348-354. Censhare, University of Minnesota:
Minneapolis, USA
Askew HR 1996 Treatment of Behaviour Problems in Dogs and Cats.
Blackwell Science: Oxford, UK
Barlow TA 2003 Hypothyroidism and behavioural change in the
domestic dog. PhD thesis, Southampton University, UK
Beerda B, Schilder MBH, van Hooff JARAM and de Vries
HW 1997 Manifestations of chronic and acute stress in dogs.
Applied Animal Behaviour Science 52: 307-319
Christiansen FO, Bakken M and Braastad BO 2001
Behavioural changes and aversive conditioning in hunting dogs by
the second-year confrontation with domestic sheep. Applied
Animal Behaviour Science 72: 131-143
Clark, GI and Boyer WN 1993 The effects of dog obedience
training and behavioural counselling upon the humancanine rela-
tionship. Applied Animal Behaviour Science 37: 147-159
© 2004 Universities Federation for Animal Welfare

Dog training methods 69
Hennessy MB, Williams MT, Miller DD, Douglas CW and
Voith VL 1998 Influence of male and female petters on plasma
cortisol and behaviour: can human interaction reduce the stress
of dogs in a public animal shelter? Applied Animal Behaviour Science
61: 63-77
Hubrecht RC 1993 A comparison of social and environmental
enrichment methods for laboratory housed dogs. Applied Animal
Behaviour Science 37: 345-361
Hubrecht RC 1995 Enrichment in puppyhood and its effects on
later behaviour of dogs. Laboratory Animal Science 45: 70-75
Jagoe JA and Serpell JA 1996 Owner characteristics and inter-
actions and the prevalence of canine behaviour problems. Applied
Animal Behaviour Science 47: 31-42
Johnston B 1995 Harnessing Thought. Queen Anne Press:
London, UK
Lieberman D 1999 Learning Behaviour and Cognition.
Wadsworth: London, UK
Marschark ED and Baenninger R 2002 Modification of
instinctive herding dog behaviour using reinforcement and punish-
ment. Anthrozöos 15: 51-68
Most K 2000 Training Dogs: A Manual. Dogwise Publishing:
Wenatchee, USA
Mugford R 1981 Problem dogs and problem owners: the behav-
iour specialist as an adjunct to veterinary practice. In: Fogle B (ed)
Interrelations Between People and Pets pp 295-318. Charles C
Thomas: Springfield, IL, USA
OFarrell V 1997 Owner attitudes and dog behaviour problems.
Applied Animal Behaviour Science 52: 205-213
Overall KL 1997 Clinical Behavioural Medicine for Small Animals.
Mosby: St Louis, MO, USA
Podberscek AL and Serpell JA 1997 Environmental influences
on the expression of aggressive behaviour in English Cocker
Spaniels. Applied Animal Behaviour Science 52: 215-227
Roll A and Unshelm J 1997 Aggressive conflicts amongst dogs
and factors affecting them. Applied Animal Behaviour Science 52:
229-242
Serpell JA 1996 Evidence for an association between pet behav-
iour and owner attachment levels. Applied Animal Behaviour Science
47: 49-60
Siegel S and Castellan NJ 1988 Non-Parametric Statistics for the
Behavioural Sciences, Edn 2. McGraw-Hill: New York, USA
Voith VL, Wright JC and Danneman PJ 1992 Is there a rela-
tionship between canine behaviour problems and spoiling activi-
ties, anthropomorphism, and obedience training? Applied Animal
Behaviour Science 34: 263-272
Wells DL and Hepper PG 2000 The influence of environmen-
tal change on the behaviour of sheltered dogs. Applied Animal
Behaviour Science 68: 151-162
Animal Welfare 2004, 13: 63-69