Dog training - use of handheld remote-controlled training devices (e-collars): report

5. Evidence Gathered

In keeping with SAWC ways of working ^[5], the Dog Training Working Group followed a three-step approach.

i. Gather written evidence from stakeholders (proponents and critics of e-collar use).

ii. Gather verbal evidence at stakeholder meetings.

iii. Conduct a literature review.

Stakeholders

A list of the organisations and individuals engaged is at Appendix I. Stakeholders were selected to achieve a balance of views from both proponents and critics of e-collar use, and because they had had previous contact with Scottish Government consultations on the topic.

Written evidence

In line with SAWC ways of working, a call for written evidence was made through a six-part questionnaire that referenced some of the Working Group's areas of concern. n line with SAWC ways of working, a call for written evidence was made through a six-part questionnaire that referenced some of the Working Group’s areas of concern. This was emailed to organisations and individuals on 02 December 2021 and results were compiled. They can be found on pages 16-20 and in Appendix II.^[6]

The questions posed by SAWC and the responses are briefly summarised and discussed below. The questions were intended to explore the rationale and methods behind e-collar use, as these are not widely discussed in the published literature. This resulted in those stakeholders that are critical of e-collar use summarising their opposition to use of the aids without elaborating on their use. Proponents of e-collar use provided answers to the questionnaire and those responses form the basis of the Q&A summary below.

Written evidence provided in response to the questionnaire can be read in full at Appendix II.

Written responses from critics of e-collar use

British Veterinary Association (BVA)

BVA provided a policy statement 'BVA and BSAVA ^[7] policy position on the use of aversive training devices in dogs and cats' (Appendix III), and a peer-reviewed paper 'Efficacy of Dog Training with and without Remote Electronic Collars vs. a Focus on Positive Reinforcement' (China et al. 2020). In summary, the BVA:

• support and recommend positive training methods as the most effective training intervention for companion animals in terms of health, welfare, and behavioural outcomes.
• believe that aversive training devices, including electric collars, used as a means of punishing, or controlling behaviour of companion animals are open to potential abuse and incorrect use, and have the potential to cause welfare and training problems.
• call for a complete ban on the sale and use of electric-pulse training collars used to deliver electric shocks in dogs.
• believe that electric-pulse collars raise several welfare issues, such as the difficulty in accurately judging the level of electric pulse to apply to a dog or cat without causing unnecessary suffering.

The paper submitted by the BVA (China et al. 2020) concludes that training with positive reinforcement was found to be more effective at addressing the study target behaviour as well as general obedience training. Further, the authors of the paper state that positive-reinforcement training poses fewer risks to dog welfare and quality of the human-dog relationship than e-collar use ^[8].

Animal Behaviour & Training Council (ABTC)

The ABTC's code of conduct 'prohibits the use of e-collars or any device that emits an aversive stimulus'. As a result, ABTC practitioners do not use e-collars, but rely on rewards-based training. ABTC asserts that the risk in using (such) 'punitive devices' is that they will aggravate the target behaviour, thus increasing the risk to all concerned. Good results can be achieved without e-collars.

The Deaf Dog Network (DDN)

The Deaf Dog Network does not support any form of physical punishment in the training of dogs. The DDN does not endorse training methods using tools such as check/chains, prong collars, squirting water, shock collars and the like, but approves and supports modern, science-based force and fear-free training methods.

Questionnaire responses from proponents of e-collar use.

The commentaries below summarise respondents' key points. Full responses can be found at Appendix II.

Q1. What behavioural challenges are e-collars used for? In your experience are cases referred by veterinary surgeons? What proportion?

A. The inclusion or not of an e-collar in training requires evaluation of the animal and its behaviour. Candidate dogs may already have been exposed to other training methods without success. The devices may be used to treat challenging behaviours, such as aggression, predation, off-lead control and pulling on the leash. In some instances, e-collars are used more routinely, for example for training and routine communication, including with deaf dogs, or to enhance response to learned commands, and aim to teach dogs confidence, relaxation, and better communication. E-collar training has also been used to assist elderly, infirm or disabled owners, as well as the physically able, to control their dogs. E-collars may also be used to train basic obedience, development of leash pressure, trick training, and agility.

Some proponents of e-collars believe that vets are not qualified to advise on dog behavioural matters, or the inclusion of an e-collar in training, and that 'For behavioural issues where welfare is concerned then you would need to consult a behaviourist as opposed to a vet. Vets have no knowledge or expertise when it comes to remote collars and dog training in general.'

Occasionally, behavioural referrals are received by e-collar-using dog trainers from vets, but not explicitly for e-collar use.

Q2. What is the range of training methods used before an e-collar is used? Is this always the process?

A. ECMA recommends that where possible a suitable, competent supervisor should be consulted prior to e-collar inclusion in a training programme to increase success rates and minimise the potential for improper use.

It is not possible to generalise about the methods used as there is much variability between e-collar users. However, an assessment of cases may be undertaken and may be 'based upon consideration of multiple factors', and is not limited simply to 'what methods have been used before'. Such assessment may or may not result in use of an e-collar.

An e-collar can be used in any type of training to increase understanding between owner and dog, and is 'layered on top of a command that a dog has already been taught to full proof that command'. E-collars have been used in dogs, which have previously been trained by positive reinforcement methods, but which were not effective.

Remote trainers (e-collars) allow pets and owners the freedom to go on walks safely, without a lead. Following a period of training, the e-collar 'allows the owner to reach out to the dog while it is in full flight and break its concentration', 'Nothing else exists that is capable of providing clear, instant, and consistent communication to the dog through associating unpleasant consequences with dangerous behaviours.'

Q3. What risk does a dog pose and / or what is the risk to the dog or others if an e-collar isn't used?

A. The ECMAresponse summarises the risks listed by the respondents:

'Risk potential is determined in accordance with social expectations, legal obligations, and the individual animal/context. It is not possible to provide a single, cover-all answer to the question asked. Typically, risks to the individual dog and other people or animals includes:

• Death of healthy dogs having been shot by a livestock keeper.
• Death of healthy dogs by a veterinarian following livestock worrying.
• Death of healthy dogs following a court destruction order under various laws, such as S3 offences under the DDA 1991, the Protection of Livestock Act 1953 and S2 Dogs Act 1871.
• Avoidable death or serious injury to livestock animals, companion animals (cats / dogs) or wildlife resulting from unresolved chasing/failing to come when called using reward-based training.
• Substantial loss of income to livestock keepers.
• Veterinary destruction of healthy dogs for unclassified 'undesirable behaviour' which has proven intractable following reward-based training.
• Dogs being surrendered / abandoned / rehomed due to resolvable behaviours such as chasing other animals, people or vehicles, reactivity / lunging on lead, failing to come when called, excessive barking or over-excitability due to prolonged lead confinement.
• Lifelong confinement to leads and inadequate exercise.
• Owner prosecution under animal control laws.'

Further comment referred to 'enabling access to quality electronic collars under supervised tuition encourages owners to come forward and self-refer, rather than remain silent and leave the risk of further attacks unresolved.'

Q4. Describe the range of equipment used, including make and model. Describe how it is used, including frequency of use, duration etc.

A. ECMA note that: 'remote trainers' (i.e., e-collars), enable dogs to go on walks safely without a lead. Following a period of training during which the dog is taught to become fluent in understanding the behaviour required to remove and avoid the stimulation from the collar, the remote trainer allows the owner to reach out to the dog while it is in full flight and break into its concentration. ECMA add that, nothing else exists that can provide clear, instant, and consistent communication to the dog through associating unpleasant consequences with dangerous behaviours.

Several brands of e-collar are referenced by stakeholders, Ecollar Technologies, Dogtra, Garmin, Sport Dog, Dogtra Arc, and Chameleon, not all of which are ECMA products. It was reported that, in general, the mini educator from Ecollar Technologies tends to be the best remote collar to use for pet dog training. Owners may learn how to use the e-collar by reference to YouTube videos.

Patricia Bowerbank gave a comprehensive example of how e-collars are used, which, over a 2-week period, broadly split into:

i. 'Tuning' dog in, or familiarising dogs to the sensation of the e-collar, walking on a 10ft lead giving occasional low-level (level 2 of 100 potential stimulus settings, with 100 being the maximum shock stimulus) exposure to the electric stimulus. This exercise would last around 2 minutes then rest, let the dog absorb for around 10 to 20 minutes then repeat, repeat about 3 times.

ii. Practice recall on 30ft long line using low-level stimulus.

iii. Basic obedience work and 'layering' with the level 2 on the e-collar. Practice until 'proofed'.

iv. Remove verbal and other stimuli and use only the e-collar stimulus.

The level of the stimulus is adjustable and can be set to a minimum level to suit the animal's nature and the situation.

Q5. How do you assess that e-collar use is necessary?

A. ECMA advises that use of e-collars should complement pre-existing training efforts, with a focus on reinforcing desired behaviours. Working with a competent supervisor greatly increases success rates, whilst minimising the potential for improper use.

Jamie Penrith advised that e-collar users must appreciate breed (traits), heritage, individual learning history, owner requirements and capabilities, choice of equipment, the context/s in which the behaviour presents and, more importantly, the probability of inclusion [of an e-collar in a training programme] improving the existing situation.

Patricia Bowerbank states that [the need for e-collar use] depends on the issue or training need. If the owner wants to stop a critical /risky behaviour that they are unable to with the leash, plethora of other tools, training, education, exercises listed above, the e collar is an option.

Q6. On what proportion of dogs that you work with are e-collars used? How many dogs do you work with annually?

A. Respondents use e-collars on approximately 20-50% of dogs, and with approximately 250-1500 dogs/year.

Additional comments / covering notes

ABTC

'It is our strongly held view that there can be no justification for subjecting dogs (or any animals) to unnecessary pain and / or stress in order to modify their behaviour. Suitably educated and assessed practitioners will only employ more reliable, humane methods to achieve more dependable results.

Any training device that works on the basis of inflicting an aversive stimulus on the dog for not carrying out the handler's wishes and / or until they exhibit acceptable behaviour compromises the animal's welfare, subverts the human / animal bond, is unethical and illegal under the terms of the Animal Health and Welfare (Scotland) Act 2006.'

Lez Graham

'Although I don't use a 'shock' collar I would hate to see them banned. There are a great many trainers that can use them with finesse and incorporate them as part of a training protocol rather than them just being placed on the dog and the dog punished for misdemeanour; these trainers are helping owners with their wayward dogs enjoy freedom on walks whilst keeping dog and other animals safe. As such I would urge instead of banning them from public sales, licencing them to trainers that are conversant in conditioning them effectively.'

Roddy Kirk

'We recognise both the efficacy of the tool and the possibility of misuse. Therefore, value the opportunity to be involved in the discussion around sensible legislation to minimise misuse. In a climate where "unwanted behaviour" is one of the primary causes for the euthanasia of dogs any tool which has a proven track record of stopping such behaviours is an important factor in stopping the need for such drastic action, not only for the welfare of the dogs themselves but also for the owners whose lives are deeply affected by trying to manage them.'

Jamie Penrith

'In closing, I would like to repeat my offer to meet with the SAWC in order to answer and clarify any and all points raised and to provide further information as required / requested. I am regarded a national lead figure on the responsible consideration / inclusion of electronic training collars for dogs, having studied the topic extensively from both theoretical and applied perspectives.'

Deaf Dog Network

'Underpinning the ethos of the group is this statement: The Deaf Dog Network does not support any form of physical punishment in the training of dogs. We do not endorse training methods using tools such as check / chains, prong collars, squirting water, shock collars and the like. We approve and support modern, science-based force and fear free training methods.'

Section Summary

In discussion, both stakeholders that support and those that oppose the use of e-collars acknowledge the risk of misuse and abuse, and the resultant dog welfare harm inherent to the devices. ECMA stipulates minimum requirements for its products, but there are e-collar brands currently in use that are not regulated by ECMA and therefore have no specific quality or electrical output requirements or limitations.

Proponents of e-collar use support their application in various training situations, including to prevent and treat livestock predation, routine use to enhance obedience training, and treatment of other unwanted behaviours and use on deaf dogs. They may apply both positive reward-based techniques, and e-collar use to the same dog, and advocate that better results may be achieved by those experienced in e-collar use.

Stakeholder meeting

The SAWC Dog Training Working Group, SG representatives and stakeholders met at the Science & Advice for Scottish Agriculture (SASA) building, Edinburgh on 14 April 2022. A summary of the evidence presented is given below. Attendees are listed at Appendix I; initials in parenthesis in this report refer to stakeholders' names.

An outdoor session at the start of day enabled the Working Group to observe dogs wearing e-collars and experience e-collar use in trained dogs in close proximity to their trainers. Three dogs, including one deaf dog, came to heel promptly with low-level shock stimulus and exhibited no outward sign of discomfort. The group was able to examine the e-collars and their control handsets.

During the meeting, both proponents and critics of e-collar use observed that their methods alone promote animal welfare.

Morning session summary

During this session the SAWC working group met proponents of e-collar use. The opinions below reflect the views expressed at the meeting by those individuals, or when in [parentheses], those expressed by SAWC members.

A video of sheep being predated and dogs in training for predation deterrence was shown by JP. Proponents of e-collar use argue that in Wales there has been a significant increase in sheep worrying since the use of e-collars in cats and dogs was banned in 2010, and state that livestock worrying by dogs is a greater animal welfare concern than the potential for misuse of an electronic training collar, which when used correctly, instils a lasting and immediate avoidance of the livestock by the dog. Further, they state that prohibition of e-collars may encourage owners to revert to other options, which are cruder and present their own animal welfare issues.

The Association of Responsible Dog Owners advocate keeping dogs on a lead when near livestock, the training of dogs to prevent predation, and after an episode of predation the use of methods, including positive reinforcement and use of e-collars in training. One justification for e-collar use is that when lead restraint is used (either routinely or to prevent predation), a large dog can cause injury to its owner when it attempts to run away, particularly if the owner is frail or weak. Anti-predation training may be used to avoid this scenario.

E-collar users state that e-collars 'can lead to negative welfare states for some dogs', but that the welfare impact of not using e-collars is greater.

The technical aspects of e-collar use were discussed. The setting (i.e., the level of shock delivered by an e-collar to the dog) used for an aversive/startling stimulus may be 65-75 out of 100, or at the maximum (i.e., 100), but the stated objective is less about the amount of pain delivered, but more the 'startling' effect, which the dog then associates with the livestock. For more routine training on an e-collar with settings from zero to 100, a setting of 2-5 may be used. E-collars with gradations up to 100 may not deliver more shock compared to those with fewer settings, but the smaller gradations may provide finer control. There is significant breed and individual variation in the level of stimulus/shock required. Minimum and maximum electric output levels are specified in ECMAstandards for the aids, but cheaper, non-ECMA-compatible collars may exceed those levels. In Australia, where some use is licensed, ECMA^-approved collars are used and comply with the ECMA-derived standards.

A paper by Elliffe, critical of China et al. (2020), was circulated prior to meeting by JP (Elliffe, 2020). The author states that due to flaws in the study, it is not possible to determine whether e-collars are less effective than positive reinforcement methods and concludes that using e-collars as negative reinforcement is problematic, and that they are more suited to positive punishment in training. [The author states that e-collars are reliable in reducing predatory behaviour in dogs, although the China et al. (2020) paper does not address this issue, but merely refers to e-collar manufacturers' claims that the aids may suppress such behaviour].

In discussion regarding whether control or regulation of e-collars was required, some e-collar users do not seek a ban or any controls, but believe that education of owners and other stakeholders is the key to better use and understanding of e-collars. Other users believe that regulation would allow e-collar use to be more openly promoted, discussed, and understood.

Proponents question whether, if regulation controlling e-collar usage was implemented, all usage would be controlled and observe that whilst discussion tends to focus on anti-predation training, '90% of the time' e-collar use is not aversive / punishing and observe that e-collars produce such a positive outcome that they should be used routinely (RK).

ECMA advised that the operating model used in Victoria State, Australia is an exemplar of how regulated use can be applied.

Critics of e-collar usage observe that regulation would be hard to enforce, but a regulatory model whereby chosen trainers have specific equipment, which they retain at the end of the training period, would be easier to regulate. [The Scottish SPCA get few complaints about e-collars, but this may be because users of e-collars are not witnessed, as they use them in the countryside. Furthermore, remote-control e-collars, anti-bark collars, and containment collars appear similar, and an observer cannot readily discern which type, regulated or not, is being worn by the dog].

Compared to scenarios where e-collars are used on daily basis, when applied for anti-predation training, their use is for a limited period, not a continued tool – 'it's a tool, not a crutch' (JP). For anti-predation training, clients are encouraged not be reliant on the collars, but it takes time to train the dog. However, periodic reinforcement training is required. Owners buy a specific product recommended by the trainer and once clients have started the training process, they usually retain the e-collar.

Improper use is a recognised risk with e-collars, for example, through not using the collar for the use it was intended, or out of anger, or anything 'that causes unnecessary pain or suffering' (AC).

Afternoon session summary

In this session SAWC met critics of e-collar use. The opinions below reflect the opinions of this group; those comments in [parentheses] reflect those of SAWC members.

The Animal Behaviour & Training Council was formed in 2010, following a report published in 2008 by CAWC that concluded that regulation of trainers and behaviourists was highly desirable and would have benefits for animal welfare (DM). Voluntary self-regulation of the sector, and assessment of individuals is now conducted. ABTC would not agree to use of electronic devices, due to animal welfare, preservation of owner-animal bond, lack of long-term viability of e-collar results as they are based on positive punishment and not a desirable way of training any animal and so those that use them would not qualify for use of them. ABTC would not deem it appropriate for e-collar users to train themselves.

ABTC does not believe that use of e-collars should be regarded with the same utilitarian Harm-Benefit Analysis as, e.g., surgical procedures carried out under the Veterinary Surgeons Act, i.e., procedures that benefit animal welfare that are carried out by select professionals, but cause pain. Rather, owners should be educated in the patient application of positive reinforcement methods, and trainers, behaviourists and others should encourage responsible training methods, rather than 'abrogating the responsibility by using e-collars' (DM). Education of owners for responsible dog ownership is needed. People want a quick fix for training their dogs, and e-collars are used for some people, who 'don't want to put the groundwork in' (DM). This quick gratification, through e-collar use for training dogs, is not to representative of progressive society that we want to be part of.

Section summary

E-collars have been used by trainers to treat dogs that are at risk of, or that have previously exhibited, livestock predation. Although often e-collar training to prevent livestock predation is initially performed by, or under direct supervision of a trainer, we were concerned regarding the continued use of e-collars by owners when unsupervised. Regulation of e-collar use is currently in place in some jurisdictions, which allows for their use under certain stipulations.

E-collar users state that e-collars 'can lead to negative welfare states for some dogs', but that the welfare impact of not using e-collars is greater. E-collar use is difficult to witness, and remote control e-collars, anti-bark collars, and containment collars appear similar - an observer cannot readily discern which type of collar is being worn by a dog, nor whether it is currently in use.

ECMA advised that the operating model used in Victoria State, Australia is an exemplar of how regulated use can be applied.

Literature review

High-quality research regarding the beneficial or adverse effects of e-collar use is limited but growing. Most of the research evidence relating to e-collars falls into the categories of background information, anecdote, expert opinion, and questionnaire responses. The evidence resulting from higher quality research, such as observational studies, comparative studies, and meta studies, is more limited. Questionnaire data from users of e-collars provide mixed results.

A significant weakness of the available literature is that studies do not compare the efficacy of available techniques, e.g., in the case of predatory behaviour, comparing reward-based training vs. e-collar training vs. environmental management.

A further weakness is that where e-collar-based schemes are currently in use, such as snake or kiwi, Apteryx spp., predation, no field evidence is available to measure the efficacy of the regimen in protection of dogs, or prey species.

E-collar devices

ECMA is the sole industry body to publish technical requirements for electronic pet training and containment collars (ECMA, 2012b). These are a voluntary standard adhered to by select manufacturers. The standards include specifications for unit energy output, output current, and peak current output. Non-ECMA-compliant versions of e-collars that can be purchased online may not have safety mechanisms, such as maximum output controls, or guidance on usage. Although some non-ECMA manufacturers may produce e-collars to ECMA-equivalent specifications, the overall lack of technical requirements for non-ECMA e-collars may present a greater welfare risk to dogs trained with such tools.

The strength of the electric stimulus produced by an e-collar can vary between models and brands of device, even when used on the same strength setting (Lines et al., 2013). The electrical resistance of a dog's skin and therefore the shock delivered can vary significantly due to factors, such as density of the hair overlaying the skin or whether the skin is wet or dry (CAWC, 2012). It has been reported that e-collars generally deliver stimuli of similar energy when the resistance is varied, but it is not known whether the perceived stimulus strength to the dog is unchanged (Lines et al., 2013).

The number, age and quality of e-collars in circulation is not known, but in one study, 170 types of e-collar were available for purchase online (AW1402, 2013), and it is estimated that between 300,000 (in England) and 560,000 (in the UK) devices are believed to be in use (Blackwell et al, 2012; Pickwick, 2014). It is not known how many e-collars ECMA members sell annually, nor the specification of all available e-collars, and it is therefore not possible to know the volume, and proportion of ECMA^-compliant and non-compliant e-collars in circulation. It is reasonable to assume that a number of non-compliant collars are in circulation, given that in our stakeholder discussions, these were being used by subject matter experts.

In its Code of Practice ECMA advises that 'all training and use of electronic collars must be done either in accordance with ECMA member's guidelines or under the close supervision of a qualified dog trainer' (ECMA, 2012a, p4). The ECMA training document describes the general usages of e-collar systems both for problem behaviours, such as chasing/predation, and for basic obedience. Further they advise that collars should not be used for certain categories of dog; dogs less than six months of age, pregnant or nursing bitches, dogs that cannot respond appropriately due to injury, illness, or age, dogs with aggressive tendencies, and dogs with certain anxiety-related disorders, such as separation anxiety (Radio Systems Corporation, 2011; ECMA, 2012a, p.4).

Purposes for which e-collars are used

The methods applied in the use of e-collars in dog training vary considerably within the community of dog trainers. There are no peer-reviewed data on how many trainers recommend the use of e-collars, or how they are used (e.g., 'communication' tool vs. positive punishment), but survey data from 2012 reported that the main reason for e-collar use was to prevent dogs from chasing livestock, or wildlife, and to improve recall (Blackwell et al., 2012).

General Obedience

Some trainers describe using e-collars at low settings as a 'communication tool' as a cue to perform a behaviour, for example, to come back to the owner to get a reinforcer / reward. Dog trainer Roddy Kirk explained in his written response to SAWC that 'most commonly a remote collar is used to communicate with the dog at a distance from their owner. The remote collar guarantees clear and concise communication even with real life distractions.' (Appendix II, Question 1). Katz observes that by using e-collars, dogs may be kept under control at a distance from their owners (Katz, 2010). A study by China et al. (2020) concluded that training with an e-collar did not provide any significant benefit over positive reinforcement when training a 'come' and 'sit' command. In one self-selected online survey, 92% of contributors stated that electronic training aids 'resolved their problem', and 99% of respondents 'state that there were no negative effects'. Respondents used electronic training aids to address a number of behavioural issues, including predatory behaviour, failure to come when called, and off-lead reliability (ARDO, 2022).

Overall the published evidence of e-collar use for 'general obedience' is limited, with studies such as China et al. (2020), omitting key details such as the shock regimen used during training. Therefore, the evidence relating to this type of use is largely anecdotal.

Deaf dogs

Whilst advocated for use by some trainers, for example, in his written response to SAWC, Dog Trainer Roddy Kirk states that 'A remote collar can be used for a variety of issues, from a deaf dog to an aggressive dog and everything in between' See Appendix II, Question 1), another stakeholder, The Deaf Dog Network, is critical of the use of e-collars in deaf dogs.

Prevention of predatory behaviour

Predatory behaviour is mostly genetically influenced, where dogs chase other animals, which can lead to injury or fatality for the animals concerned (Miklósi, 2014). Attacks on sheep and other livestock by dogs is a worldwide concern ^[9]. Work commissioned by the Scottish Government in 2019 evidenced that 14% of sheep farmers reported that dogs had attacked or chased their sheep in the previous 12 months, and estimated that the total number of incidents of dogs, chasing or attacking sheep in Scotland in the period 1 May 2018 to 30 April 2019, was around 7,000, with the true figure likely to be within the range of around 4,500 to 10,000 (Gov.scot, 2019).

There is some evidence in the published literature that aversive training, using an e-collar, can be effective in preventing predatory behaviour in dogs under specific circumstances. For example, training dogs to avoid taxidermy kiwis in New Zealand ^[10] (Dale et al., 2017, Dale et al., 2013), reducing the probability of sheep attacks in Norway (Christiansen et al., 2001) and approaches to sheep in the UK (Cooper et al., 2014). Authors have remarked that using this approach to control predatory behaviour requires good timing on behalf of the trainer/handler. Poor association between the stimulus and cue can be ineffective at changing behaviour (Hiby et al., 2004), has negative welfare consequences (Schalke et al., 2007), and there is an increased risk of fearful and aggressive responses by the dog (Polsky, 1994, Christiansen et al., 2001).

Breeds may respond variably to e-collar-based sheep aversive training, with Elkhounds requiring more electric shocks in one study (Christiansen, 2001), and Alaskan Huskies not responding readily to the treatment (Hansen, 1997). These authors did not compare the efficacy of e-collars to other methods.

Howell and Bennett (2020) surveyed and interviewed dog training and behaviour experts, and stated that in relation to dogs with a history of predatory behaviour, experts who used only positive, reward-based techniques were 'typically (but not exclusively) more pessimistic about preventing predatory behaviour than experts who incorporate both aversive and reward-based methods into their training practices'. However, experts concluded that effectively managing a dog's environment, such that it never has an opportunity to engage in predatory behaviour, is the best way to prevent the behaviour (Howell and Bennett, 2020).

E-collar devices have been tested on predatory coyotes, Canis latrans, and wolves, C. lupus. In coyotes, maximal-strength electric shocks delivered by observers was found to reduce the majority of attacks on enclosed lambs – and the use of e-collars 'would likely have somewhat limited application in the field' and would need to be tested 'under field conditions' (Andelt, 1999). Studies have also been carried out with e-collars on captive wolves, e.g., where they were able to stop a wolf's approach to a chosen food but did not elicit a deterrent effect when wolves were repeatedly offered food. Authors wrote that these findings highlighted 'the complexity of application of nonlethal techniques in real-world situations' (Shivik, 2003).

Duration of e-collar use in training is variable. For dogs undergoing predation behaviour modification, these appear to be intermittent sessions, in between which the owner keeps the device to be used as trained by the dog trainer, but in their absence. For more general routine communication, to supplement the learning process, owners retain the e-collar for use on a continuing basis.

Use of e-collars internationally

1. Kiwi aversion training in New Zealand

In New Zealand the Department of Conservation (DOC) is charged with conserving the country's natural and historic heritage. Kiwi conservation falls under the DOC wildlife protection remit. The threat to these flightless birds from pets is well recognised, and advice to dog owners in kiwi-populated areas includes that dogs should be contained on owners' property, should be on a lead when out walking, and that Kiwi Avoidance Training (KAT), using e-collars, may be stipulated for hunting dogs, and farm dogs intending to access to some areas (Kiwi Avoidance Training, 2022). Successful completion of avian aversion training does not override any dog access restrictions for an area and is just one of many tools used to reduce the threat pets pose to ground-nesting birds.

Dogs should first be trained to obey basic obedience commands. They are then fitted with an e-collar and exposed to dead ground-dwelling native birds and their faeces in a controlled situation. Whilst sniffing these, the dog is given a 'short sharp shock' through the e-collar. Once the dog shows good kiwi avoidance behaviour, a certificate is issued. Refresher training is held every six, 12 or 24 months, depending on the outcome of the first training, to make sure the dog remembers what it has learnt (Kiwi Avoidance Training, 2022).

Dale (2017) observed that in KAT training, many dogs displayed avoidance to the training stimuli after the first training session, and that 100% of dogs responded by the fifth session. Some dogs displayed lower levels of avoidance, including older dogs trained for the first time and those with a three-year gap or longer between training sessions.

The authors did not present data on the technique's effectiveness in the field, or efficacy versus other techniques, such as positive reinforcement, and recommended that the ecological translation of the training be investigated – i.e., to establish if KAT training benefits kiwis in their habitat (Dale, 2017).

2. Snake avoidance training in Australia

The licensed use of e-collars in the State of Victoria, Australia is enabled by the Prevention of Cruelty to Animals Regulations 2019, made under section 42 of the Prevention of Cruelty to Animals Act 1986. The regulations specify by whom, under which circumstance, and which types of e-collars may be used and supplied. For example:

Requirements for use of authorised electronic collars

A person must not use an authorised electronic collar on a dog or cat unless— (a) a veterinary practitioner has examined the physical health and temperament of the dog or cat and reasonably believes that the dog or cat is suitable to have an authorised electronic collar used on it; and (b) the dog or cat is over six months of age; and (c) a collar is not left on the dog or cat for more than 12 hours in any 24-hour period; and (d) the use is in accordance with any instructions for use of the collar provided by the manufacturer; and (e) the dog or cat is introduced to the use of the collar in accordance with a training programme that complies with the Victorian Code of Practice for the Training of Dogs and Cats to Wear Electronic Collars; and (f) the design and technical specifications of the collar comply with standards that have been approved by the Minister. (POCTA, 2019, Part 2, Division 2, section 24)

Snake avoidance training is an example of aversive training, to which e-collar use is applied, and ECMA (in a written response to SAWC) states that: 'In the State of Victoria (Australia), a tried and tested welfare-focussed regulatory model has been in place for over 10 years, having been agreed between ECMA™ and relevant stakeholders. This model involves veterinary input. It was recently reviewed (2020) and deemed fit-for-purpose.'

In the USA, snake avoidance training, using positive reinforcement (clicker) training, has also been applied (Karen Pryor , 2022).

E-collar use vs. keeping dog on lead where there is a risk of predatory behaviour

Many of those concerned about livestock predation observe that the most essential management method is to keep dogs at risk of undesirable behaviours, such as predation, on a lead, or safely enclosed. In 2022, the UK Minister of State, Department for Environment, Food and Rural Affairs, stated that;

'Dog owners can prevent incidents of livestock worrying through keeping their dogs on a lead in the vicinity of livestock and/or undertaking appropriate training. It is important that dogs are trained to behave well, ideally from a young age, and introduced gradually and positively to different environments, people and animals. Reward-based training for dogs is widely regarded as the preferred method of training. Owners who have concerns about controlling their dog's behaviour may take advice from their vet or a suitably qualified dog behaviourist or trainer. The Animal Behaviour and Training Council maintains national registers of appropriately qualified trainers and behaviourists.'(UK Parliament, 2022)

In one police report, 89% of dogs that killed livestock were unaccompanied by their owners (therefore had escaped, or been let off the lead), and 5% of offences were caused by repeat-offending owners/walkers (North Wales Police, 2022). This suggests that secure dog enclosures and public education may also be required to prevent predation of livestock by dogs.

Section summary

We acknowledge that e-collars have been shown to prevent predatory behaviour in dogs under specific circumstances, but in those studies there are no robust data comparing the efficacy of e-collars to other methods, nor the technique's efficacy in field conditions. The use of a lead and safe enclosures to control dogs can also effectively prevent livestock predation without the welfare harms associated with aversive methods. Although e-collars can also be used for general obedience training, there is evidence that positive reinforcement methods are equally as impactful and are not associated with the same risk of welfare harms.

Owing to variability between breed responses, and individual dogs' susceptibility to electric shock, judgement is required in the use of e-collars. ECMA advises that 'all training and use of electronic collars must be done either in accordance with ECMA member's guidelines or under the close supervision of a qualified dog trainer'.

E-collars and Animal Welfare

The ethical cost of effectiveness

Operant conditioning is a method of learning, in which a behaviour becomes stronger or weaker depending on its consequences (Chance, 2003), and which forms the basis of the two broad categories of training methods applied in dog behaviour modification scenarios. 'Aversive based training methods' (including e-collars) rely for their effect on positive punishment and negative reinforcement and 'reward-based training methods' rely on of positive reinforcement and negative punishment (Blackwell et al., 2012). When e-collars are used to modify behaviour, they are typically used as positive punishment, but may be used as negative reinforcement, by applying the shock stimulus until the dog does the required behaviour (Masson 2018). A third category of training, 'balanced' training, incorporates both reward-based and aversive training methods, and is referenced by some proponents of e-collar use (Hunter et al., 2020, ABDT, 2017).

Ethological (i.e., behavioural) algorithms have been referred to in the literature, describing behaviour-change strategies applied to both humans and animals. The application of these algorithms is intended to highlight that behaviour change should be achieved with the minimum of aversive effects. Further, they are a reminder that effectiveness is not the only measure by which a training intervention should be measured, and that 'if we make effectiveness the only criterion by which we determine the appropriateness of an intervention, we risk failing to consider some other ethical objectives' (O'Heare 2012). Friedman (2009) also observes that an ethical hierarchy of behaviour-change procedures encourages actions that are both effective and humane'; Fig.1 below shows the series of steps in which behaviour-change procedures should be considered and applied (taken from Friedman, 2009). The authors highlight many detrimental potential side effects of punishment, including increased aggression, generalised fear, apathy, and escape avoidance fears (Friedman, 2009)

The ECMA Code of practice aims 'to specify the minimum standards required when using an electronic collar on dogs or cats' (ECMA, 2012a), and to follow the LIEBI (Least Intrusive, Effective Behaviour Intervention), and LIMA (Least Intrusive, Minimally Aversive) algorithms regarding intensity of electric stimulation delivered by e-collars. Users should use the minimum stimulation necessary to achieve the training objective' (ECMA 2012a.).

Pain and distress

Prevention of pain and distress is a well described goal of good animal welfare, for example, in the Five Freedoms formula (Brambell, 1965). All mammals are known to be capable of experiencing emotions such as pain and distress.

Behavioural indicators, suggesting that dogs are in a negative welfare state during training with e-collars, include incidences of vocalisation (Tortora, 1983, Schilder and van der Borg, 2004, Salgirli et al., 2012), low body posture (Schilder and van der Borg, 2004, Beerda et al., 1998), backwards ear position (Salgirli et al., 2012) and time spent in a tense state (Cooper et al., 2014). Studies that have shown success in e-collar use, also reported that e-collar use caused signs consistent with discomfort, for example, vocalisation, head shaking and jumping (Christiansen et al, 2001, Christiansen et al, 2001a and Salgirli, 2012). Further, in another study, 36% of owners reported their dog vocalised on first use (a strong implication the dog experience distress and/or pain) and 26% on subsequent use (AW1402, 2013).

Shocks that cause pain can interfere with the dog's ability to learn (Polsky, 1994), and if a dog's motivation to continue the behaviour is strong, shock may require repeated application (Polsky, 2000). Polsky further states that pain induced by e-collars may subsequently lead to aggression (Polsky, 2000).

Physiological stress

In a study to assess stress levels by measuring salivary cortisol in laboratory Beagles exposed to electric shocks via an e-collar, those dogs which were able to clearly associate the electric stimulus with their action did not show considerable or persistent stress indicators, but dogs that received unpredictable shocks did show severe and persistent stress symptoms (Schalke, 2007). As a result, Schalke states that 'use of these devices should be restricted with proof of theoretical and practical qualification required and then the use of these devices should only be allowed in strictly specified situations' (Schalke, 2007, p.379).

Salgirli et al. (2012) reported an increase in cortisol with e-collar use, but more dogs showed elevated cortisol with a verbal quitting signal. Steiss et al. (2007), Beerda et al. (1998) and Cooper et al. (2014) found no increase in cortisol levels with e-collar use. Although these studies indicate that the behaviour of dogs can be affected by e-collar use, there are important limitations to the utility of these studies. These include, but are not limited to, complicated experimental designs with inadequate controls, selective or unjustified data analysis and over-extrapolation of results.

Whilst adverse welfare effects, such as vocalisation, have been observed in acute response to e-collar activation, there is conflicting evidence on the long-term welfare implications of e-collar use. Owner questionnaire data showed that e-collar use had no negative effect on their dog's behaviour during the year after its use (Christiansen et al., 2001) and similarly, cognitive-bias tests did not demonstrate any long-term effects on affective state of dogs due to e-collar use (AW1402, 2013).

Conversely, behavioural observations from training German Shepherd dogs for use as guard dogs described increased incidences of stress-related behaviours in dogs trained with e-collars compared to controls. This difference was recorded outside the training environment, and the authors concluded that the use of an e-collar may influence the dog's long-term wellbeing in a negative way (Schilder and van der Borg, 2004).

Unintended Behavioural change

The results of several questionnaires indicate that the use of training methods, based on positive punishment and negative reinforcement (not exclusively e-collar based training), are related to higher incidences of behaviour problems, aggression, and fear (Blackwell et al., 2008, Casey et al., 2014, Herron et al., 2009, Hiby et al., 2004). The results of the self-selected ARDO study (ARDO, 2022), referenced earlier, suggests that many e-collar users do not report such issues. Rooney and Cowan (2011) observed that there are clear links between a dog's current behaviour and its owner's reported training history, as well as with the owner's present behaviour, with aversive training methods impacting negatively on future behaviour. Chavez (2012) states that aversive training methods compromise dog welfare, and that high levels of punishment may have adverse effects on a dog's future behaviour, whereas reward-based training may improve a dog's subsequent ability to learn (Rooney and Cowan, 2011).

Observational studies have shown that dogs trained with aversive methods showed stress-related behaviours in training (Cooper, 2014; Haverbeke, 2008; Schilder 2004).

Recent studies have found that the use of aversive training methods, including leash jerks with choke or pinch collars, e-collars, water pistols, rattle cans, citronella spray collars, slapping the dog, yelling at the dog and leaning towards the dog in a threatening way.may induce longer-term negative mood states (Vieira de Castro et al., 2020), and that dogs appear to show a more pessimistic cognitive bias, where trainers report using two or more aversive training methods (Casey et al., 2021). These studies observed the use of a range of training tools. Therefore, although the results of Vieira de Castro (2020) and Casey (2021) highlight the welfare impact of aversive training methods in general, care must be taken when specifying individual training tools/methods such as e-collars.

Authors have observed that the use of positive punishment in the form of a pinch collar or an electronic collar can have detrimental effects on dogs' physical and mental welfare, and that using punishment without clear or consistent instructions of what is expected of the dogs can lead to fear and stress (Ziv, 2017).

German Shepherds trained with e-collars were shown to display more behavioural signs of fear towards their handlers, even outwith training environments, compared to those trained without shock collars (Schilder and van der Borg, 2004). It has also been argued that aversive stimuli could be associated with an unintended object or person (Polsky, 1994). As an example, Polsky (2000) reported a case in which a dog fitted with a boundary-fence-associated electronic collar, associated a person walking near the electronic fence with the pain from the shock and exhibited human-directed aggression. However, as noted by CAWC (2012), a major limitation of this report is that it lacks comparison with other examples of apparently unprovoked attacks by dogs that do not involve electronic boundary systems.

Physical harm

ECMA identifies the risk of pressure necrosis (i.e., death of the skin on the neck) from e-collar use, especially in dogs who wear e-collars for prolonged periods: 'If electronic collars are worn by dogs for prolonged periods, pressure from the dermal contacts can reduce blood supply to the skin resulting in skin damage.' (ECMA 2012a). Instructions for electronic collar use include advice regarding the risk of pressure necrosis (ECMA 2012b). The risk of pressure necrosis is also reported anecdotally, but not in the peer-reviewed literature, which suggests either that the risk is exaggerated, or that such injuries are underreported to vets.

Despite the warnings given by ECMA that prolonged wearing of e-collars carries the risk of skin damage, there is very limited published evidence that wearing an e-collar can cause physical injury to a dog. Polsky (1994) argues that electronic collars must be 'snug' to have effective contact with the skin, and that lesions can develop from abrasion of the electrodes, which are worse in animals with skin conditions. An online questionnaire of dog owners in France reported that 7% of e-collar users reported physical damage/burns (Masson et al., 2018). However, other authors refute that e-collars cause physical damage (Klein, 2000, Lindsay, 2005).

Effect of trainer competence

The ECMA Code of Practice states that all training and use of electronic collars must be done either in accordance with ECMA members' guidelines or under the close supervision of a qualified dog trainer (ECMA, 2012a). Further, problem behaviours associated with training are more likely to arise because of inconsistency and inappropriate delivery of both unpleasant stimulation and reward. These are issues which need to be addressed by better informing and instructing trainers regardless of the training technique they employ (ECMA, 2012a).

The ECMA Code of Practice also states that e-collars must not be used when the dog is out of sight until the trainer is confident that the dog will return to them. Using e-collars when the dog is out of sight risks mistiming of electric stimulation via the e-collar and the undesirable behaviour. The need for users to adequately observe behaviour prior to application of shock is encompassed in CAWC's observation that 'it is clear that poor contingency between the application of an electrical stimulus and the behaviour to be modified can give rise to both behaviour and welfare problems.' (CAWC, 2012). The inappropriate use of e-collars carries the risk that the dog may not link delivery of the e-collar stimulus with the conditioning stimuli (Klein, 2000), and that reinforcement may lead to stress and pain (Schilder, 2004).

Because some interventions can exacerbate unwanted behaviours and, for example, render dogs more aggressive, care is required when deciding to whom we entrust the training of our pets and the methods that are to be used (Chavez, 2012).

As noted above, in one study Christiansen observed success in prevention of dogs approaching or attacking sheep. However, he recommended that e-collars be used only for the purposes described in the paper, and only if it is used by skilled trainers with special competence in dog behaviour, learning mechanisms, and of the particular device used (Christiansen, 2001).

Efficacy as training device

As stated above, the efficacy of training methods should be considered against their superiority to other methods, and their welfare impact on the dog. Considering these factors, the European Society of Veterinary Clinical Ethology concludes that better training options than e-collars exist and that it 'strongly opposes the use of e-collars in dog training and urges all European countries to take an interest in and position on this welfare matter' (Masson, 2018).

DEFRA states that whilst it has been suggested that e-collars might have a beneficial impact in preventing dangerous and harmful behaviour by dogs, which are out of control, and e-collars might be proposed as a last-resort measure for poorly behaved dogs, which would otherwise be put down, little evidence has been provided to support these suggestions. On the other hand, evidence about the harm e-collars inflict on pets has been growing (DEFRA, 2018). In the same vein, Cooper states that there is no consistent benefit to be gained from e-collar training, but there are greater welfare concerns compared with positive reward-based training (Cooper, 2014). Ziv (2017) holds a similar position, stating that there is no evidence to suggest that aversive training methods are more effective than reward-based training methods.

Survey data from 3,897 dog owners demonstrated that the use of reward-based training techniques had higher effectiveness in teaching a dog recall or preventing chasing than those who used e-collar training (Blackwell et al., 2012), but this could be due to confounding factors, such as the seriousness of the problem behaviour and owners' perception of success (Ziv, 2017).

In the context of routine education, rather than modification of unwanted behaviour 42 police dogs (Salgirli et al., 2012) were trained to maintain a 'heel' position with distractions. The effectiveness of an e-collar, prong collar, or a quitting signal (that was conditioned to signify the withdrawal of a reward) was evaluated. A greater number of dogs learned to disregard the distraction with the use of the electronic collar (39) and the pinch collar (32), compared to only three dogs with the use of a quitting signal. The authors propose that the e-collar was the most effective form of punishment. Ziv (2017) suggests that that the dogs receiving the quitting signal did not understand what was expected of them in this specific setting.

Results from owner questionnaires reveal that highest obedience scores were reported by owners who used reward-based training only, followed by those who used a combination of reward- and punishment-based methods, and lastly by those using punishment only (Hiby et al., 2004). However, this result is based on subjective answers to questionnaires and causality cannot be drawn. Conversely in a questionnaire given to owners seeking advice from a dog behavioural service, owners reported positive effects of aversive methods, such as lead corrections, prong collars, e-collars and forcing a dog to lie down using a leash (Herron et al., 2009). The authors highlight some limitations of the study, in addition to those applicable to all owner questionnaire methodology, and highlight that correction or punishment alone does not selectively reinforce desirable behaviour and is an inefficient way to train an animal to perform a specific behaviour.

Pet dogs, trained with reward-based methods, have been reported to perform better in novel training tasks (Rooney and Cowan, 2011) and cognitive-bias tasks (Vieira de Castro et al., 2020), whilst military dogs that received more aversive stimuli during obedience exercises and protection work were more distracted and showed poorer performance compared to dogs that received less-aversive stimuli (Haverbeke et al., 2008).

A study conducted by China et al. (2020), re-analysing video originally recorded in previous DEFRA-funded project AW1402, concluded that training with an e-collar did not provide any significant benefit over positive reinforcement alone, when training a 'come' or a 'sit' command. In this study 63 dogs with owners that reported off-lead problem behaviours, such as poor recall, were allocated to three training groups, each run by a professional dog trainer. One group used only an e-collar, one group used dummy e-collars and the last group used predominantly positive reinforcement. Measures of training efficacy included the number of commands given to elicit the response, and response latency. Dogs trained with positive reinforcement only had reduced response latency to commands, required fewer commands and had a better response after a single instruction. No difference was found in owner perception of training efficacy between the groups (Cooper et al. (2014). The research also found that many owners did not read the e-collar manufacturers' instructions prior to use. However, there has been criticism of this work, noting design constraints, inappropriate analysis of results, and that the results are not generalisable to the use of e-collars for all training purposes (Sargisson and McLean, 2021, Bailey, 2020, Elliffe, 2020).

Section Summary

Given the data available, our main concern is for the welfare of dogs being trained by e-collars and the pain caused by the device. This can have detrimental side effects, such as increased aggression, and generalised fear and apathy. Although adverse welfare effects have been reported in the acute response to e-collar activation, we acknowledge that there is conflicting evidence regarding long-term welfare harms associated with their use.

There is evidence that in many circumstances, e-collars are not as effective as other methods used in dog training. SAWC acknowledges that e-collars may be effective in certain circumstances, to prevent predatory behaviour, but notwithstanding the avoidable risk of welfare harm and lack of justification for their use when compared with other methods, the operator must be competent in their use in order to produce the desired behaviour change.

Currently, with no regulation of who can access and use e-collars, there is an increased risk that e-collars may be used ineffectively, increasing the potential for welfare harms.