In Chapter 9 of Stephen Lindsay's HANDBOOK OF APPLIED DOG BEHAVIOR AND TRAINING, Lindsay reviews the research on electric collars for training. His chapter is illuminating. He's a renowned positive trainer who advocates minimally aversive, minimally intrusive, relationship-based training. After reviewing the research, he concludes that, used correctly with modern low-stim techniques, e-collars can be substantially less aversive than gentle leaders and other head collars, and that they can be very effective for many training situations. I was surprised, to say the least, but he's an excellent trainer and scientist, and I really respect his work. We've used an e-collar on Vanya (when he began chasing the neighbor's cattle, after we invested in thousands of dollars of double-fencing). I was very reluctant to go that route, but for my dog anyway, it results in much less stress than being trained with a flat collar or a gentle leader. Modern e-collar training works on a low-stim negative reinforcement model, and the dog learns that she can turn off the stim by offering a particular behavior (the stim is aversive, but at a very, very low level--it's mildly annoying).
As another well-known positive trainer named Greta writes about Low-stim negative reinforcement training:
"Plenty of bird dog training is done this way and it involves training the dog that he can turn off the very low stim level by choosing a behavior or obeying a cue. The trainer starts with the dial turned down very low, and then turns it up one notch at a time until she sees the dog react just a little, e.g. stop what he is doing and look perplexed. Once this level is established (which looks like an equivalent of a human hearing an unfamiliar noise or noticing a slight itch and trying to figure out where it is coming from), a behavior is elicited and when the dog does the behavior, the stim is turned off. Once the dog understands how to turn off the stim, the behavior is shaped in higher and higher distraction environments. At first the dog is escaping — the stim is turned on, and then turned off when the dog follows the cue. After that, the dog is avoiding — the stim is never turned on since the dog is following the cue promptly. For a while, if the dog doesn’t follow the cue promptly, there will be a nick — a very short burst of stim — to remind the dog that it should avoid, i.e. perform the cue promptly. When distractions get very big, yes, the dial will be turned up, and pain may or may not be inflicted depending on the dog, the trainer, the distraction… but this shouldn’t happen many times. (And, in my limited exposure, the dog isn’t necessarily feeling it a whole lot more than he felt those initial little test stims — instead, the level is higher because on those occasions when it “needs” to be turned up, the dog is higher on adrenaline and is not feeling small sensations at all anyway — this is a physiologic state that has been studied a whole lot. So the perceived discomfort may not be so great anyway.)
This can work great with very high drive dogs, and it works through conditioning.
After all, people say that clicker training a recall can never work with their dog because there is no food in the world that can compete with chasing a deer. And what do clicker trainers say to this? They say “but you don’t just wave food at the dog when he is chasing a deer. You teach him in low-distraction situations that it’s always rewarding to come running when called, until it is a reflex and that reflexive learning takes over even when the dog sees a deer.” We don’t claim that food can, in fact, with no training, outcompete a deer. And the same can be true with LSR- collar training — it’s not the strength of the current that “works” when there’s a deer, it’s all the prior training that produces a reflexive turn-and-run-to-handler, that can kick in even when there is a deer out there." (
http://melissa-c-alexander.com/2010/08/ ... /#comments)
My experience with Vanya correlates closely with what Greta describes. I'd say that 99% of my training with Vanya is based on the two quadrants familiar to positive trainers: rewards for the behavior I want, withdrawal of attention or reinforcement when he offers behaviors I don't want. But I have found that an e-collar is much more effective for certain behaviors that are important for Vanya's long-term health. Ie, with an e-collar, I can allow him off-leash freedom on our 20 fenced acres, and I can take him off-leash in the hunting lands that surround our farm. Without the e-collar, he wouldn't have those options available.