Debunking "ecollar studies" - Janeen McMurtrie "See no evil, read no evil, cite no evil"

Liesje

If the dog is in a "level 1" just acknowledging another dog, why would he need to be corrected for that?

Well we are talking about dog- aggressive dogs, if the dog becomes aggressive towards 100% of the dogs he sees i dont know why you should let him get fixated to a point where you are going to have 3 times more trouble "bringing him back". You know where he is going to end. Maybe level 1 of aggression sounds more accurate but some dogs go from acknowledging to aggression in less than a second so it might depend on the dog how soon you should apply the correction. The body language will give the queue

I might be a little bit off but i would believe is the same principle as a leash correction. Maybe miranadobe could explain you a little bit more on the timing since she has more experience on the filed.

I asked because you brought up the issue of timing, about not allowing a dog to become fixated (regardless of method), which I agree with, but then you mentioned using an e-collar when the dog is as "level 1".  I'm asking you to please explain.  What is "level 1" and how would you use an e-collar?  Are you saying you would begin applying corrections at "level 1" to avoid the dog from escalating?  How do you define level 1?  It seemed like you were saying the dog is simply acknowledging the other dog (like glancing over at it and back, noticing it is there but not fixating or reacting), which to me seems very appropriate and normal and doesn't warrant being corrected.  I'm just trying to understand why you made the distinction in the context of using an e-collar...

espencer

Maybe level 1 of aggression sounds more accurate

Espencer, how many ways do I need to ask it....please define "level 1 aggression" or refer to some site or book or something so we know what you are talking about...  and you still have not addressed how you would use an e-collar in this context.  At this moment I have absolutely no idea what you are trying to say or what point is being made...

Liesje

What I don't get about these studies and discussions is that they always seem to center around using an e-collar on a dog that has problems or is aggressive.  IMO that's just asking for it.  To me the e-collar is a training tool.  Slapping it on an aggressive dog and dialing up the stim is NOT training, it's just stupid, zapping a dog into submission.  I would never judge the value of a tool based on the most stupid of uses, sort of like I don't rag on the Easy Walk harness b/c I've seen some person jerking their confused dog around with one.

I guess I am lucky I've never seen anyone try to use an e-collar in that way.

While I don't espouse their use, that was a very sensible statement, in general, about not judging tools by the "most stupid of uses."

Unfortunately, most people who use them are not as well versed as Kim is in basic neuroscience, and they often do not remember that it is not WE who decide what is or is not aversive, it's the subject organism, in this case the dog.  Timing or not, some dogs will consider a shock, however slight, to be an aversive.  As we all should know by now, that places the shock in the category of positive punishment, and withdrawal of the shock in the category of negative reinforcement, but certainly not in the category of positive reinforcement.  If you are going to make the argument that either of those quadrants will work to train a dog (name your behavior, whether it's aversion to rattlesnakes, or doing handstands), then why would you think that the positive reinforcement quadrant could not work equally as well?

spiritdogs

Unfortunately, most people who use them are not as well versed as Kim is in basic neuroscience, and they often do not remember that it is not WE who decide what is or is not aversive, it's the subject organism, in this case the dog.

Well, now....LOL. I'm not an expert by any means. I just find the brain fascinating so choose to study it. I'm going to address another point entirely, and that is.....in the case of serious aggression, and where the stim from an ecollar causes the dog's overarousal to go over the top and reach the explosion it can, chances are the dog does not even actually find the ecollar aversive at that point. The dog doesn't even likely remember the ecollar stim afterward, and at an extreme point, it may not actually have an effect on the future behaviour of the dog (aka no actual learning takes place at all). What happens is the e-collar stim at that moment can cause the adrenaline dump, which in turn causes the reaction. This isn't even a matter of learning, but a matter of the body reacting to physiological changes in the body. The dog actually has no control at this point, and learning doesn't even factor in.

I agree that the best trainers, no matter the methodology, would not let it go that far before intervening. They start far earlier, at the first sign of a dog acknowledging another dog. The problem I see therein is that I can't comprehend why any trainer would "correct" (apply punishment) to that very first sign, which would be simply a glance at another dog. At this point - the point a good trainer would notice the dog noticing the provocative stimulus - the dog is doing nothing wrong whatsoever, so I can't imagine punishing a dog for "noticing" behaviour. What a good trainer would do is acknowledge the calm behaviour and heavily reward it, thereby creating a counter-conditioning experience so the dog is left with relaxed, happy emotions when it looks at the other dog. The reward could come in the form of toys, food, and even increasing distance and then starting again - but regardless, the dog is rewarded for the appropriate behaviour rather than corrected for what I would think is a desirable behaviour - calm acknowledgement of a stimulus.

miranadobe

corvus

I have personally seen a dog that reacted to the application of an aversive whilst being aggressive become more aggressive.

You don't say ecollar, so I assume it was something else.  Using an ecollar does not always mean an application of an aversive. 

Then why use an ecollar at all? If it's the training and not the tool, then why train with an ecollar at all? Presumably you could get the same effect with a conditioned positive interruptor, for example. What about Leslie McDevitt's Look At That game? She uses a bridge as the interruptor. I've tried it with my dogs and it amazes me how fast my Lapphund can come out of a fixated state if he hears that bridge. For a dog that has gone glassy-eyed it's incredible. Works every time. And he's nowhere near as alert and clued in as my other dog.

Kim_MacMillan

I agree that the best trainers, no matter the methodology, would not let it go that far before intervening. They start far earlier, at the first sign of a dog acknowledging another dog. The problem I see therein is that I can't comprehend why any trainer would "correct" (apply punishment) to that very first sign, which would be simply a glance at another dog. At this point - the point a good trainer would notice the dog noticing the provocative stimulus - the dog is doing nothing wrong whatsoever, so I can't imagine punishing a dog for "noticing" behaviour. What a good trainer would do is acknowledge the calm behaviour and heavily reward it, thereby creating a counter-conditioning experience so the dog is left with relaxed, happy emotions when it looks at the other dog. The reward could come in the form of toys, food, and even increasing distance and then starting again - but regardless, the dog is rewarded for the appropriate behaviour rather than corrected for what I would think is a desirable behaviour - calm acknowledgement of a stimulus.

Exactly, and I think that goes to Corvus' point, too.  Why would you correct a dog for simply noticing a stimulus?  In fact, if you can time your reinforcement to coincide with that, and not let the dog get to the point of *reacting* to the stimulus, you've got it half licked right there.  If you can manage that, you never even need to get to the point of the dog ever being out of control.  We call it "working under threshold" - a concept that some trainers miss.  But, I think it's much clumsier to wait for an unwanted behavior so that you can correct it.  The exquisite finesse in training comes from the *human* noticing what the dog is about to do, then stepping in to elicit an alternative, or to reward the "notice" and not the "reaction."

corvus

Then why use an ecollar at all?

From the perspective of SchH, it is most often used b/c it allows one to add extreme distances and maintain safety.  Often it is there as a "backup" tool.  My friend's dog wears one and the helper is the person that holds the remote.  If for whatever reason the dog is out of control, he has a way to get through to the dog without a physical conflict (like a fight between a dog and person).  I've never actually seen it used this way because I've never seen our dogs get out of control, but I've seen dogs wear the collar for this reason.

Also many people use it to give more accurate corrections, both timing and the level of correction.  This is assuming one trains with corrections (another topic of debate, but whether you like it or not you have to accept that people do it).  Some of the collars have like 150 levels plus tone and vibration, continuous vs. nick, etc.  There's no way a human can consistently give over 150 different levels of correction, and be consistent with timing and reward.

A few people also use them to mark behaviors at greater distance using a tone, vibration, or lowest level stim.  It is imprinted the same way you imprint and charge a clicker.  This is far more rare but I have seen it used this way on a dog that is very "operant" and completely understands this type of mark/reward training with both the collar and the clicker.

Then there's just good old fashioned negative reinforcement.  Again if one was to train this way, the e-collar is more consistent and fair than using a prong or choker.

spiritdogs

Why would you correct a dog for simply noticing a stimulus?  In fact, if you can time your reinforcement to coincide with that, and not let the dog get to the point of *reacting* to the stimulus, you've got it half licked right there.  If you can manage that, you never even need to get to the point of the dog ever being out of control.  We call it "working under threshold" - a concept that some trainers miss.  But, I think it's much clumsier to wait for an unwanted behavior so that you can correct it.  The exquisite finesse in training comes from the *human* noticing what the dog is about to do, then stepping in to elicit an alternative, or to reward the "notice" and not the "reaction."

I agree, which is why I asked espencer to clarify because I don't understand why an e-collar would ever be needed for a dog-aggressive dog that is only acknowledging other dogs.  If the dog is fixating or escalating higher than that, the failure is on the part of the handler by not setting the dog up to succeed (either by having too little distance, too much distraction, or too little proofing of an incompatible behavior).

Thanks Liesje. I understand that there are a lot of uses for e-collars and situations in which they might be useful, but for the purposes of this discussion about using them in behaviour modification, I don't think it necessarily falls into any of those categories you mentioned. If, as was said, the training is so good, then why couldn't you achieve the behaviour modification with the vibration setting alone, for example? Someone recently told me that some dogs find the vibration more aversive than the stim, but that aside.

Liesje

e-collar would ever be needed for a dog-aggressive dog that is only acknowledging other dogs. 

espencer

level 1 of aggression sounds more accurate

espencer

Liesje

e-collar would ever be needed for a dog-aggressive dog that is only acknowledging other dogs. 

 

espencer

level 1 of aggression sounds more accurate

 

I'm not sure you have addressed Liesje's questions, but I have one of my own.  Since you think you can identify "level 1 of aggression," exactly what are your criteria for determining whether a dog is genuinely aggressive (at any level), reactive, or just over-excited to get to the other dog?  And what is so criminal about a dog, even if they are any of the above, just acknowledging the presence of the stimulus, in this case another dog?

spiritdogs

And what is so criminal about a dog, even if they are any of the above, just acknowledging the presence of the stimulus, in this case another dog?

did you read my last 3 posts?

espencer

spiritdogs

And what is so criminal about a dog, even if they are any of the above, just acknowledging the presence of the stimulus, in this case another dog?

did you read my last 3 posts?

 

Yes, and I still don't understand why you would wait until a dog does go over threshold, rather than simply proactively rewarding the dog who exhibits the "notice" of another dog.  If you intervene immediately with a reinforcement while the dog is noticing, you are classically conditioning the dog to understand that the appearance of the other dog is *good* and not bad.  If you want to talk about timing, my question would be why wait until the dog makes a mistake so that you can correct him?  Why not just time your reinforcement to coincide with his first glimpse of the other dog?  Incidentally, people have had great success with the "open bar, closed bar" exercise, which doesn't really require the dog to be "good" - the aim is, over time, to change the emotional response by feeding when the dog sees the other dog, and stopping feeding when the other dog disappears.  If the dog's emotional response changes, the behavior changes, too.