I've written a lot recently on my view of human and dog interaction and I've skirted around the old way of thinking. Yesterday I was faced with a somewhat awkward conversation as I was on the phone with a trained that preferred the archaic compulsion technique of training. He told me that operant conditioning just didn't work on some breeds of dogs and that you had to gain their respect through punishment. While I do believe that negative stimulus is required for full spectrum learning I do not think that the negative stimulus needs to be all that harsh or extreme and just like the positive stimulus it needs a strong marker. For the vast majority of pet dogs hat should mean nothing more than an unpleasant noise or a squirt in the face from a water bottle simultaneously as a loud and stern, "No."
The words no and wrong are very valuable tools in training a dog as if they don't know if what they are doing is incorrect how will they know to stop doing it?
As for the efficacy of shock or e-collars. Recent studies have shown they are no better than reinforcement techniques in training dogs. The reason they might seem more effective to trainers that don't know any better or haven't bothered to do the research into operant conditioning (which is surprising to me as BF Skinner published The Behavior of Organisms in 1938) is that the marker is immediate with the reinforcement. The dog hears a beep, gets shocked. It is more difficult to have a marker that strong and in connection with positive reinforcement. This is ho clicker training came about or why you say, "Good boy," to your dog when he sits. These are markers and should proceed closely to the reinforcement that is being used. It is also important to train a dog to a marker. That means when training begins just feed your dog treats and say good boy or click a clicker when you do it. They will start to associate the marker with the reinforcer and then they will be easier to train.
It is my opinion that the idea of punishment based training goes back to a time when we need to master the beasts of the world, and in most cases most punishment based trainers come from the working dog world, but even that has been changing recently. Think of positive reinforcement this way. It is a more true form of learning. I am of an age where we memorized the multiplication table and while that was an effective way to pass a test it wasn't a good form of learning. Imagine instead of being taught that 3 x 3 = 9 you are taught that 3 x 3 = 3 + 3 + 3 = 9. Now you could substitute any number in for the three and know that the equation for multiplication doesn't change. Add the first number together the amount of times the second number dictates and now you can do any multiplication problem and not just the ones memorized. It is true learning.
The use of strong markers and positive reinforcement in dog training is going to give you a companion and not a slave. We want to have a companion animal and not a subordinate. i used before the analogy of viewing our human/dog relationship much as we would our human/human relationships and while their might be occasions when we'd like to punish our spouses with a shock collar or chock chain it would be a quick way to end up in prison. Coercion training is the path to mastery of your dog while operant conditioning is the path to having a true companion animal.
I'll end this by quoting a published coercive trainer and let you interpret the words how you want and decide if this sounds like a good way to train a dog or a good way to get bit.
Interject your physicality on them to say ‘Hey there’s a new sheriff in town.’