Does science support balanced training?
I see a lot of social media posts about balanced training and science. And a lot of responses claiming “it doesn’t matter if science supports it what matters is that it works.”, “I don’t care about science, I care about results,”, etc. But here’s the thing. You SHOULD care about science. If science says a certain method of dog training does not work and is harmful to dogs, you SHOULD NOT USE THAT METHOD. That’s the beauty of science: it’s an organized, falsifiable method of telling us what works, what doesn’t, and what’s flat out WRONG. But an important question to ask ourselves, is what is science, anyway? One article is not science. Any article that presumes to state independently, based on one finding, if a tool or method is effective or not, can be presumed to be poorly thought out. And it HAS to be peer-reviewed, and ideally written by someone who has some sort of real credential. Pro tip: Psychology Today articles do not fall under these criteria. Google Scholar is a great place to start a search of your own.
So, if a certain method of dog training works, the science will show that, too. Let’s take a look at balanced training.
Article One: https://psycnet.apa.org/record/1926-01096-001, A note on the relative value of punishment and reward as motives, Hoge, M. A., & Stocking, R. J. This article suggests that animals learn best with a combination of punishment AND reward (go figure). This is a fastest and clearest way for an animal to pick up a new expectation. Interestingly, the researchers also observe that without combination of both methods, learning actually occurred faster with just punishment rather than just reward (not that you should ever be using just punishment in the scope of a dog’s life, but, pretty neat).
Article (presentation) Two: https://hiddenfence.com.au/wp-content/uploads/2016/06/Comparison-of-stress-and-learning-effects-of-three-different-training-methods-in-dogs.pdf, Comparison of stress and learning effects of three different training methods in dogs, E. Schalke, Y. Salgirli, I. Böhm, S. Ott, & H. Hackbarth. This piece is amazing, definitely my favourite. These researchers systematically compare the stress caused by three different training methods in dogs. They use e-collars, pinch collars, and a pure-positive method that involves a preconditioned quitting signal (self-inhibition). They found that the pinch and e-collar worked to have the dogs successfully complete the task (not breaking a heel while being distracted by another person) while the quitting signal did not. But here’s where things get interesting. Although the pinch and e-collars evoked body postures and vocalizations in the dogs suggesting submission/pain, salivary cortisol levels were highest when using the quitting signal. That is, the dogs were MORE stressed by the frustration of exhibiting self-inhibition than they were by the positive punishment induced in other methods. Which method is more harmful to dogs now, PP?
Article Three: https://www.tandfonline.com/doi/abs/10.2752/089279302786992685, Modification of instinctive herding dog behavior using reinforcement and punishment, Eve D. Marschark & Ronald Baenninger. This paper looks at working herding dogs around sheep. It concludes that although PP is a great tool to be used independently to teach certain behaviours, negative reinforcement and positive punishment are great additions to the arsenal, particularly when combatting “instinctive motor patterns”, ex. chasing wild animals. Imagine that, you should use the whole quadrant.
Article Four: https://psycnet.apa.org/record/1954-02233-001, Traumatic avoidance learning: the outcomes of several extinction procedures with dogs, Solomon, R. L., Kamin, L. J., & Wynne, L. C. In this article, the researchers show that dogs will stop performing a previously rewarded behaviour most effectively when punishment (along with other motivators) is involved, and that simply cutting off the reward associated with the behaviour (a PP method of training) is not enough.
And so, as you can see, balanced training is indeed science-based. This certainly isn’t all that’s out there, just a little mouth-wetter to get you started. Now I won’t argue the evidence isn’t limited, it is – because many ethics committees that involve members of the community (those most prone to PP nonsense) don’t approve the experiments using balanced methods as ethical, and this is a huge reason advocacy for our methods is so important. BUT, SCIENCE is still SCIENCE. And you won’t find peer-reviewed articles that blanket-deny the value of certain methods if those methods, in reality, work. So what about all these claims of pure positive training being science-based/modern? Let’s take a look at some of the articles that might suggest this.
PP Article 1: http://dogscouts.org/base/tonto-site/uploads/2014/10/620_art_training_methods.pdf, Dog training methods: their use, effectiveness and interaction with behaviour and welfare, EF Hiby, NJ Rooney, & JWS Bradshaw. This article correlates the training methods used by people to their dogs’ obedience. A fun saying in the science world is correlation does not equal causation. Ice cream purchases are highly correlated to shark attacks, but that doesn’t mean the ice cream causes shark attacks, nor that people who were just attacked by a shark have a huge ice cream craving. Likewise, there are many explanations for the correlation reported here. It could easily be the case that since training has historically made use of a lot of (often too much) punishment, new and inexperienced dog handlers tend to try these methods first, whereas people who consider themselves more modern do a few quick google searches on how to train a dog, and are bombarded with posts warning to never raise your voice at your dog. And so the correlation could have nothing to do with the actual training methods, but rather the skill level of the person using them, ex. their ability to read the dog and respond accordingly. Just one other possibility.
PP Article 2: https://www.sciencedirect.com/science/article/abs/pii/S0168159108003717, Survey of the use and outcome of confrontational and non-confrontational training methods in client-owned dogs showing undesired behaviors, Meghan E.Herron, Frances S. Shofer, & Ilana R.Reisner. Another survey, more correlations. Authors conclude, “confrontational methods applied by dog owners before their pets were presented for a behavior consultation were associated with aggressive responses in many cases.” Perhaps this speaks to the importance of not using “confrontational” methods before you seek guidance/help from someone who knows behaviour—akin to someone getting cut with a chainsaw because they didn’t read the manual. But that’s just another correlation. 😉
PP Article 3: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0102722#s2 , The welfare consequences and efficacy of training pet dogs with remote electronic training collars in comparison to reward based training, Jonathan J. Cooper, Nina Cracknell, Jessica Hardiman, Hannah Wright, & Daniel Mills. Finally. A real article. Let’s get started. I’d like to point out right off the bat the title of this paper specifies *pet* dogs. It doesn’t speak to working dogs. This article is looking at dogs referred for recall issues, livestock chasing issues, etc. that are trained by e-collar trainers with or without e-collars, and some other group that doesn’t use e-collars. The preliminary results can be completely thrown out the window because of the statement, “During preliminary studies there were negative changes in dogs' behaviour on application of electric stimuli, and elevated cortisol post-stimulation. These dogs had generally experienced high intensity stimuli without pre-warning cues during training.” Yeah, throwing an e-collar on a dog and slamming them at high levels is generally a bad idea. Go figure. Let’s jump to the main study in which, “trainers used lower settings with a pre-warning function”. Terrible little science-shame in this particular article, if you just read the abstract, which note, is all that most of the other links I’ve provided allow to the general public, it looks like a pretty damn good paper showing e-collars are bad and PP is the way to go. Shit, I guess we’re all assholes. But luckily, this particular paper is fully accessible. So let’s use that to our advantage, and take a peak at the methods. ONE e-collar trainer used a method, “broadly similar to that advocated by collar manufacturers”. This dude trained ONE dog. The other three trainers (we’ll excuse the fact that a sample size of 4 trainers is pretty poor; regardless of how many dogs), used a, “different approach”. All but one of these others failed to test the dog’s sensitivity completely and opted to just jump right in. But as you’re about to read, there was really no point to testing the dog’s sensitivity anyway, given the methods that follow. “Thereafter, for all but one dog (which was exposed to a setting at the higher end of available range) the trainers selected the highest setting available on the device and dogs were allowed to roam off-lead in a field, where sheep were present. If dogs approached sheep, then the trainer would apply an e-collar stimulus using the high setting with timings of their choice. These trainers stated that they aimed to associate proximity to or orientation towards sheep with the e-stimulus, and consequently did not plan to use pre warning cues such as the collar mounted tone or vibration stimuli as a predictor of electric stimulation.” Wow, so you throw a weird new tool on a dog and zap the shit out of them, no pretraining whatsoever, and surprise! This works, but it makes the dogs really fucking upset. E-collars are clearly unethical, let’s ban them. This article is a bunch of PP supporters making a COMPLETELY bias and, CRUEL experiment which actually should not have passed through ethics, in which they recruit 4 “trainers” (they probably- HOPEFULLY, couldn’t find any more nutjobs to abuse dogs) that obviously have no idea what they’re doing or worse no care for the dog’s welfare, to mistreat dogs on tape. Then, they fail to mention this abuse in the abstract. Go figure. Paper continues to state, “One dog referred for sheep chasing did not approach sheep during the training sessions, but received two stimuli at points when it was orientated towards nearby sheep.” Yada, yada, more blatant animal abuse and a conclusion they abused dogs. Incredible finding. A final note for this article is that again, the authors refer to the dogs as “pet” dogs. If we look further into this, we find breeds used listed as gundogs, cross breeds, pastoral and working. This is awfully vague. Now, it could be the case that most of these dogs were not people-pleasers, but because breed specifics were not included, I have a sneaking suspicion that many of these dogs were border collie types. The point I’m getting at here is of course not all dogs ever need an e-collar, and you can get by just fine without one – WITH SOME DOGS. But certain breeds aren’t going to recall off something like sheep without them. In other words, the PP methods were successful in THIS paper, but WHAT WERE THE BREEDS?
PP Article 4: https://www.sciencedirect.com/science/article/abs/pii/S1558787817300357, The effects of using aversive training methods in dogs—A review, Gal Ziv. Right away I want you to notice this is a review. Speaking very generally, reviews are useless for this kind of exploration. The kind of things we just talked about in Article 3 can VERY easily be left out of a review, i.e. the things that change a paper from really convincing to abuse, at the author’s discretion. This goes both ways. Someone advocating balanced training could just as easily right a review saying e-collars work every time and hide every paper saying they damage dogs. Without knowing the METHODS of each study reviewed, the use of the review dwindles. It also states they use surveys – how much of what they reviewed was surveys? That could also bias their results toward correlational (bad) data. I’m not saying reviews can’t be useful, they can be, but when looking at such a controversial topic it would be incredibly difficult to write a totally unbiased review. Think of it this way- reviews are a generalization of a bunch of studies, but here we want DETAILS, not GENERALIZATIONS, because look what generalizations get us (PP Article 3). The authors also state, “although positive punishment can be effective, there is no evidence that it is more effective than positive reinforcement–based training”. Well, look up, we just found some in articles supporting balanced training, and I’m pretty sure that article was quite old. I’ve also seen a lot of pro-PP articles in which they’re trying to teach a new behaviour, ex. sit. Yeah, no kidding, yelling at/shocking/hitting a dog is not a great way to encourage a behaviour. I don’t think a single balanced trainer (probably not even those crazy zap guys) suggested it was.
In conclusion, science IS important. And as a balanced trainer you should have some idea of WHY you’re doing what you’re doing, beyond simple anecdotal observation. Do your research, and think critically about what you read. If something says you’re awful, think, “why?” Is it because you’re awful? If so, stop being awful. Or is it because it’s a poorly designed study? If so, educate (POLITELY).
​