I have yet to understand how it is that people can believe that an e-collar MUST hurt to be effective when they know that a whistle or a shout (or even a gentle tap on the shoulder to attract the dog's attention) works WITHOUT hurting.
So this is where you do need to look at evidence and science to understand how the ecollar works
Ok fact dogs learn by four ways.
Positive punishment
Positive reinforcement
Negative punishment
Negative reinforcement
That is undisputed. Science trainers time all shows this is now animals learn.
So ecollars has to fall into one of those categories or else it is just working by magic.
Positive punishment is adding punishment when a behaviour is performed eg shocking a dog when they chase sheep. They get a shock when they run after sheep they no longer run after sheep. The shock has to hurt as it is adding the punishment (eg the positive bit = add)
Negative punishment is taking away something from your dog.eg they pull on the lead so you stop walking and they can not go forward. They are barking at a toy you remove the toy.
Positive reinforcement is giving a pleasant reinforcer when the behaviour is performed so dog does not chase sheep they get sausage. Everytime a dog sits it gets a reinforcer. Everytime the dog looks at you instead of the sheep it gets a reinforcer
Negative reinforcement is removing the adversive when the dog does something eg gundog trainers squeeze a dogs ear to get them to release the dummy (cant believe that they still do this!)
Where you are getting confused with comparing a whistle or a shout with an ecollar is that the whistle is just the cue. You whistle dog hears whistle and gets treat. So wherever the dog is hears whistle and returns. The dog has learnt whistle means reinforcer. If the whistle was proofed enough it may be enough to prevent chase. However in a prey driven dog this is not always enough.
The ecollar adds (positive) something unpleasant (punishment) and the pain stops the dog running off after the sheep as they are fearful they will get the pain.