Introduction
This is a companion post to Ivan Balabanov’s Training Without Conflict Podcast Episode: The Real Facts About Science-Based Dog Training.
The podcast episode can be found on YouTube. It is also embedded at the bottom of this article. This episode is for anyone who wants to understand the true science of dog training. This is a straightforward overview of the peer-reviewed and validated learning science on this topic over the last 50+ years.
This article is not a replacement for watching (or listening) to this podcast. This article only provides all referenced studies, with a quick summary statement to their relevance. Most of the text is from the slides provided by Ivan in the podcast video.
I compiled these references as Ivan’s podcast is important for advancing the art of dog training. While I did the legwork on making the materials more easily accessible. The full credit for outlining the science and finding the relevant studies and materials goes to Ivan Balabanov. I encourage anyone who wants to advance their understanding of dog training to attend Ivan’s Training without Conflict Certification Course. As a TWC-Certified Trainer, I can truly say it is worth every penny.
Strawman Arguments
This podcast and the presented studies are inconvenient for the force-free community, their organizations and the veterinary groups, peddling feel-good, misinformation and lying to the public. Some know what they are doing but many simply don’t know how wrong they are in their positions. The following quote captures this well:
One of the great challenges in life is knowing enough to think you’re right, but not knowing enough to know you’re wrong.
– Neil deGrasse Tyson
Some try to brush aside the studies presented in Ivan’s podcast because they are older. But, them being older doesn’t mean they are outdated. They are older because this understanding of learning science has existed for decades. This information isn’t new; it is established. We also don’t cast aside Newton’s law of gravity because it was discovered in 1665. The same applies to these studies.
These studies are not just peer-reviewed. All of them have been reaffirmed in subsequent experiments many times. None of them have ever been contradicted. They are an essential part of science-based dog training. That they don’t fit the predominant narrative or upset ideological bias is unfortunate … for people who don’t care about the facts.
The Problems of the Force-Free Studies
In contrast, most of what the force-free community presents are studies based on surveys and questionnaires. Peer-reviewing those, means to only review the methodology of the survey process. The survey results can’t be peer-reviewed. They will change with every new group of people interviewed. That hardly compares to scientific experiments. Surveys and questionnaires also can’t be retested or reproduced unless you survey the same people with the same questionnaire twice. It should be pretty obvious that that wouldn’t make much sense.
Anyone who has ever followed an election poll knows that surveys and questionnaires are just guesses and assumptions. One reason is that many people will interpret the questions differently, making the results suspect. Further, how you ask a question also changes the answers. There are many more reasons surveys and questionnaires aren’t worth much.
If any force-free trainer could directly contradict any of these findings, they would have by now. But they haven’t been able to … in over 50 years. That alone makes it pretty obvious where the truth lies.
How to Use This Article
There is a lot to read here, as the scientific evidence Ivan presented is overwhelming in favor of his arguments. Start with what interests you the most. I tried to make this reference page as easy to use as possible. All study sections are collapsed by default. If you click on “Review the Science” you will see all studies relevant for each area. The main sections have time codes that correspond to the podcast and the order of everything is the same as in the podcast as well. Happy learning!
References
- Book: Punished by Rewards
-
Position Statement on Humane Dog Training from the AVSAB:
Claim:
The application of aversive methods – which, by definition, rely on application of force, pain, or emotional or physical discomfort – should not be used in canine training or for the treatment of behavioral disorders.
Claim:Training Effectiveness: Reward-based training methods have been shown to be more effective than aversive methods. Multiple survey studies have shown higher obedience in dogs trained with reward-based methods. Hiby et al. (2004) found that obedience levels were highest for dogs trained exclusively with reward-based methods and lowest for dogs trained exclusively with aversive-based methods. Dogs trained with a combination of rewards and aversive-based methods (often referred to as ‘balanced’ in the dog training industry) produced lower obedience levels than reward-based but better than exclusively aversive-based training. Aversive training has been shown to impair dogs’ ability to learn new tasks.
The Above Statements Are Simply False
This has been documented in 100s of scientific papers.
In some cases punishment might be the ONLY option to stop unacceptable and/or dangerous behaviors to self or others.
In this article we review the actual science and some of the most important scientific papers on learning science. This information is not new and has been known for over 50 years. Certain groups just don’t like these facts and want to cast them aside. But, that is not how science works. We don’t just ignore facts because we dislike them. All of the studies on this page are peer-reviewed, validated multiple times by different scientists and have NEVER been refuted or proven incorrect EVER! These studies represent the current state of knowledge. None of these are based on questionaires or surveys, which by definition can’t be independently validated. All of these are based on actual, scientific experiments and have been independently validated many times over.
Claim: Negative Punishment is better than Positive Punishment
Negative Punishment (withholding a reward) increases the stress in a dog significantly more than Positive Punishment with a prong or e-collar. Negative Punishment is more stressful for dogs than other forms of punishment.
Review the Science
Claim: Timeouts Are Better Than Positive Punishment
In 11 studies, people preferred giving themselves electroshocks instead of just being bored for 6 to 15 minutes. Social isolation is one of the most severe forms of punishment for humans there is. However, it doesn’t work well in all cases and it doesn’t work well in dog training.
Non-contingent release from time out may be a critical factor leading to negative associated effects.
Review the Science
Is There a Place for Aversives in Dog Training?
Although positive reciprocity (reciprocal altruism) has been a focus of interest in evolutionary biology, negative reciprocity (retaliatory infliction of fitness reduction) has been largely ignored. In social animals, retaliatory aggression is common, individuals often punish other group members that infringe their interests, and punishment can cause subordinates to desist from behaviour likely to reduce the fitness of dominant animals. Punishing strategies are used to establish and maintain dominance relationships, to discourage parasites and cheats, to discipline offspring or prospective sexual partners and to maintain cooperative behaviour.
Review the Science
Common Arguments Against Punishment
1. Claim: Punishment Induces Emotional Changes that Interfere with Learning
That would only be true when done incorrectly but also applies to positive reinforcement, or any other approach when performed incorrectly. “It’s harmful when done wrong” is not a valid argument.
Studies show the effectiveness of positive punishment in reducing problem behaviors tends to be associated with a wealth of positive side effects.
Review the Science
The positive side effects tend to far outnumber any negative side effects associated with positive punishment (contingent shock).
Review the Science
Many studies have found that using punishment as a behavior modification technique may also increase the incidents of wanted behaviors.
Review the Science
2. Claim: Punishment Provokes Aggression
Aggression responses are not restricted to Positive and Negative Punishment. Aggression responses can also occur with Positive Reinforcement when fading out food (extinction).
Experiments have shown that parts of Positive Reinforcement are aversive as well and will lead to aggression. The transition from food—Positive Reinforcement—to extinction is an aversive event. Aggression is sometimes a major side effect of that extinction.
Review the Science
Fortunately, subsequent work has suggested that the problem of elicited aggression is not really serious in most situations. Because aggression can easily be suppressed through the use of contingent punishment.
Review the Science
Further experiments, found that near-zero levels of elicited aggression could be produced by punishing each attack, even when non-contingent shocks were scheduled every 30 seconds during 2-hour sessions.
Review the Science
3. Claim: Positive Punishment is Ineffective
Punishment can be highly effective for the treatment of a variety of behavior disorders. Punishment has also been proven to be more effective in many instances than treatment with positive reinforcement techniques or extinction.
Review the Science
Although the use of punishment has been controversial for number of years, research findings suggest that punishment still remains an important option for behavior problems.
Review the Science
The use of punishment is especially effective when the reinforcers maintaining the problem behavior can’t be identified and/or controlled.
Review the Science
Punishment is the treatment of choice for life threatening behavior that must be suppressed rapidly to prevent serious harm to itself or others.
Review the Science
4. Claim: Punishment Does Not Teach A Desired Behavior
Punishment Does Not Teach A Desired Behavior.
This is a silly argument. Punishment is not supposed to teach dogs new behaviors or what to do instead of a problem behavior. That’s what Positive and Negative Reinforcement is for. Punishment is used to reduce and extinguish behaviors.
5. Claim: Punishment Creates Fear
This is misleading as fear of dangerous, injurous or destructive behaviors is a good thing. Appropriate fears keep all of us alive. For example, if we can teach a dog to be afraid of rattlesnakes, it will avoid them and not get killed. Fear of rattlesnakes is healthy. Fear of dangerous things is one of the most natural ways of learning not to do something that will harm or kill us (or a dog). This would be similar with car chasing, livestock predation, self-mutilation or other dangerous or unacceptable behaviors. In many instances, appropriate fear can be a powerful motivator for learning and can help dogs adopt good behaviors that keep them safe.
6. Claim: Aversives Create Learned Helplessness
When dogs learn how to successfully escape an aversive, they will show normal learning behavior of escape and avoidance learning. There is no negative fallout and they develop resistance to learned helplessness. Without pre-training on how to avoid aversives successfully, learned helplessness will develop. So, the key is proper pre-training.
When dogs are initially exposed to an inescapable electrical stimulus, they will not try to learn an avoidance strategy later, even when one was available. This is referred to as learned helplessness.
This claim is correct but represents the worst-case scenario and requires the trainer to be incompetent. The same researchers found in a subsequent study—two years later—that pre-training with an electric stimulus created resistance to the development of learned helplessness.
Review the Science
More recent studies found that exposure to aversives that can be controlled by the animal’s behavior help build increased resiliance not only to the aversive in question but more broadly to stressors in general. Therefore—theoretically—controlled exposure to aversives could somewhat paradoxically improve the long-term well-being of dogs.
Review the Science
Negative Reinforcement
Every reinforcement includes both negative and positive aspects. It is impossible to separate positive reinforcement from negative reinforcement.
Review the Science
The results of the analysis suggest that contingent skin shock is by far the most effective procedure and has the most favorable side effect profile.
Review the Science
Training Equipment
Force-Free trainers state that head collars and no-pull harnesses are the only training tools that should be used as they are not aversive and don’t cause pain.
Head collars and no-pull harnesses are aversive tools. Additionally, these collars are alleged to work by mimicking natural dominance behaviors thereby increasing deference and obedience to the owner.
Review the Science
The Electric Collar
… led to the conclusion that animals, which were able to clearly associate the electric stimulus with their action, i.e. touching the prey, and consequently were able to predict and control the stressor, did not show considerable or persistent stress indicators.
Review the Science
All had experienced years of failed attempts to control their aggression through large pharmacological interventions and restricting their freedom of movement via restrictive environments. The use of comprehensive multifaceted behavioral programs involving punishment resulted in dramatic and lasting reductions in aggression, the elimination or great reduction of drug use, and major lifestyle improvements.
Review the Science
In a study of 41 Elkhounds, 29 hunting dogs and 68 English Setters behaviors indicative of motivation for chasing or attacking sheep were examined in three different way and successfully suppressed.
Review the Science
There is a controversy about the use of aversive stimuli for the treatment of self-injurious, aggressive, and other dangerous behaviors with individuals with mental retardation and autism. The validity of argumentation in favor of CS is, amongst others, related to safety of the electrical shock (i.e., given its effectiveness). This study adds to this argument by demonstrating that 48 (healthy) individuals who were administered a total of 480 clinical electrical shocks failed to experience any negative effects (e.g., fainting, aggressive reactions, escape).
Review the Science
When treatment was compared to baseline measures, results showed that with all behavior categories, individuals either significantly improved, or did not show any change. Negative side effects failed to be found in this study.
Review the Science
The study ‘Evaluation of an Aversion-Based Program Designed to Reduce Predation of Native Birds by Dogs’ is the most comprehensive study on supressing predation through aversion training ever conducted. The study spanned multiple years from 1998-2007. During this time frame a totals of six training sessions were conducted. After the fifth training session all of the dogs showed 100% avoidance towards the Kiwi birds. This result remained the same during the sixth training session. Of the 1156 dog in this study, 843 dogs stopped chasing Kiwi after only one shock. 313 dogs needed two shocks to stop. 128 needed three shocks to stop. Only 42 dogs needed four shocks to stop completely. No dog needed more than four shocks. The study included follow-ups to assess long-term effects. No negative long-term side effects were found.
Review the Science
The following three studies are a great example on how the force-free community tries to manipulate studies to create the findings they want. Unfortunately for them, they were caught in this case.
This poorly constructed study is quoted a lot by the force-free training community. The focus of the training was on training recalls from chasing sheep. However, the dogs were kept on leash and the sheep in play pens. Read that last sentence again. Yes, that is what they did to prove rewards-based training works and e-collars are unnecessary. Hardly a valid setup. Thanks to Sargisson and McLean this was called out a year later (see below).
Rewards-based training is effective in stopping dogs from livestock predation. Shock-collars are unnecessary.
Review the Science
Fortunately, in 2021 Sargisson et al, reviewed the work of China et al, 2020 and stated: Attacks by dogs are unlikely to be initiated when the owner is within 1m (3ft) and the dog is on-lead. China et al.’s results shed no light on the possible behavior of the dog off-lead or when the owner is absent, and therefore cannot be used as an empirical justification for removing e-collars as a technique for treating dogs with behavioural problems.
Review the Science
The output of an e-collar tested by the New Zealand Department of Scientific and Industrial Research (Dix, 1991) was found to produce 3000 times less electrical energy than that allowed by standards for electrical fences, six times less electrical energy than that produced by the static discharge produced by walking on a carpet, and 50 times less than what is considered necessary to reach pain thresholds.
Review the Science
The traditional measure of intensity dynamic range is the ratio of the stimulation currents required to produce sensation threshold Is and pain threshold Ip. Ip/ls typically ranges from 2-4 for unexperienced subjects and 6-8 for experienced subjects. This means when someone is conditioned to the sensation it takes about double the intensity to get a reaction.
