Do Confrontational Dog Training Methods Really Work?

I have seen that study as well, and I think it's very interesting that people report success with the alpha roll.  It's hard to argue that a dog is probably temporarily stunned by having that happen, and stops the behavior that precipitated it at that time, which is, I suspect, why people think it's successful.  The Cattle Dog I recently helped re home is now in a home (with a lovely positive trainer who decided not to foster, but to permanently keep him) where he does very well and is occasionally handled by her five year old with great success.  He is a dog that is reliable in public, goes to charity walks and day care, loves people and is friendly to other dogs even when he is on lead.

At his other home, where the son in law repeatedly alpha rolled him at the age of five months, and was receiving little mental stimulation in addition to the several mile walks, he was offering lots of aggressive displaced herding behavior.  When he finally left a huge black and blue on the owner, he was taken to a behaviorist who suggested the "safer" methods of dealing with him, but they didn't even attempt to follow through (they were already afraid of him) and he got dumped at the shelter.  Mind you, this dog had awesome obedience skills - I used to use him as a demo dog at class because he was so smart and such a fast learner (and the person who owned him did not report the aggression to me until he was ready to dump the dog).  Clearly, not the right home for a sensitive dog with a work ethic and a sense of fairness...but a very clear example of how well the change to a positive scenario can help a dog fit in to the niche that humans would like him to be in.

 I think the method really depends on the dog and possibly the breed. I prefer non-confrontational training methods though.

I can't think of one single reason why anyone would think an alpha roll is a decent training tool.  Besides the fact that yes, I agree it's blatantly confrontational, there is no way to use it correctly no matter what method you use.  I cannot be used as -R.  It won't work as +P because in order to time it correctly you'd already have to have your hands on the dog.  Not to mention it's not an appropriate correction for anything. It is an incomplete process because the dog cannot be expected to recover, offer a desired behavior, and earn the reward.

I also laugh at the idea that 48% is a "success rate".  IMO, things don't work for two reasons, either it's the wrong method or the person is doing it wrong.  So according to the article, over half the people using alpha rolls shouldn't be using them in the first place (DUH!) and/or are doing them wrong (double duh!).  To me a dog understands something when he'll do it at least 85% of the time.  The other 15% quickly comes with repetitions and proofing.  Even the most complicated, or stressful, or dangerous things I've trained my dogs have been in no more than three training sessions of no more than a few minutes at a time.accomplished.  So not only is 48% another obvious DUH-why-would-you-do-this, but I'd like to know how they even quantify the "successes".

My one critique is that the synopsis offers quite a range of "confrontational" methods and doesn't seem to really define or qualify the term.  I know of some dogs that have come up the leash at their owners after being nagged repeatedly on a regular collar.  I think in most cases, the "confrontation" depends more on how a tool is used and interpreted by the dog than what the tool actually is or how it is used ideally.  Another example is how people approach dogs.  Just this weekend I saw a 7 year old girl jump up to a dog and reach to pet it, only intending to be nice but to the dog it was a confrontation and she snapped and recoiled away from the girl.  Depending on the dog, you can be more confrontational trying to pet the dog than using a prong collar.

Actually, the one thing I have a problem with in all such studies is that they are based on client reports, not on the effects seen by a trained observer, or in a double blind study.  An owner may think the technique has been effective, when it has only produced learned helplessness.  Or, an owner may simply have pushed the behavior down, only for it to erupt with greater force (and often more damage) later.  So, I think the most that can be said is that there seems to be a *correlation* between confrontational methods and aggression.  I do agree with Liesje that dogs' perception is important, since different dogs find different things more or less confrontational.  But, I would not argue that the fact that a dog sees a frontal approach as disconcerting a justification for the use of prong collars;-)  In fact, even among knowledgeable trainers who still use prong collars, there is some agreement that they are not appropriate on aggressive dogs whose behavior could be exacerbated by the infliction of a painful stimulus at the moment when they see another dog, for example, lest they become even more aggressive each time a novel dog appears.

I agree with that criticism as well.  And my comment is not to "justify" prong collars (FWIW, many people do use prong collars to intentionally ramp up the dog's drive and aggression).  As I have said in every thread on this topic, I do not make decisions on what tools I used based on other people's idiocy and misuse.  I can find logical reasons for using just about every tool and method BUT alpha rolls.  I just found it interesting that they are treating alpha rolls and squirt bottles in the same category.  To me the word "confrontational" kind of implies you are ready to physically pick a fight with your dog.  I don't see how any good can come of that regardless of whether your are flipping your dog, popping his collar, staring him down, etc.

Liesje

To me the word "confrontational" kind of implies you are ready to physically pick a fight with your dog.  I don't see how any good can come of that regardless of whether your are flipping your dog, popping his collar, staring him down, etc.

My thoughts exactly. 

 Us folks in the positive training camp don't deny that you can produce results with 'confrontational' methods. Do alpha rolls suppress behavior in some dogs? Of course. Aversives and punishment (whether it's an alpha roll or a leash pop) supress all behavior - not *just* the behavior being punished.

Aversive stimulation tends to suppress behaviors and not just the discrete behaviors you are attempting to punish but also all behaviors. Using aversive stimulation is fraught with insidious consequences and these can only interfere with training. Remember, ‘aversive’ refers to stimulation that you act to escape and/or avoid. That means it is unpleasant. Just think about your own experience. Do you think people making things unpleasant for you really promotes an environment in which you work to your highest potential? It may indeed reduce a particular behavior but there will be other effects won't there? We know in the work world that managing by fear is not the best way to get the best out of your employees and the same goes for dogs. If you realize that reinforcers (pleasant things) drive behaviors, and you control those (as opposed to unpleasant things, which we simply learn to work around) then you are on the right track. If you want a behavior, make it worth their while, and if you want someone to like you, make interacting with you pleasant, not unpleasant. If you don't like a behavior, make some other behavior in its place more worth their while, while making the problem behavior less effective. Aversives result in aggression, emotionality, disempowerment and other difficult to predict problems. This cannot influence your training positively.From http://www.associationofanimalbehaviorprofessionals.com/whats_wrong_with_dominance.html

So really, alpha rolls and other confrontational and scary methods work by creating an environment where the dog is afraid to do things. They may no longer do that annoying behavior you corrected him for, but he'll probably stop doing other things as well. At least for a short while. Punishment is an incomplete program - it merely tells the dog what he should NOT do (as that thing gets him 'Bad Things for Dogs';) but does not tell him any appropriate behaviors to perform instead. As punishment does not cause a behavior to become extinct, the behavior will eventually pop back up unless an incompatible behavior is trained in the interim. When that happens, owner usually up the punishment, and the cycle continues. And to top it off, punishment causes learning to become more difficult - animals (both human, and dog) do not learn well when aversives are used. So it makes it more difficult and harder for a dog to learn what is the appropriate action to take.

Also, aversives and punishment cause a dog to associate you with bad things. I don't want my dog to think I produce Bad Things for Dogs. Why would I want him to be afraid of me? Why wouldn't I instead want to create a positive environment where he feels safe and comfortable with me, where he knows I won't (because I never have) yell at him, hurt him, or scare him?

Sure, aversives can produce results. They can suppress behaviors you don't like (and ones you do!), but we have better, safer, more humane options out there. Why result to causing a dog discomfort, pain, or fear, when we do not have to? 

There is also the issue that the conclusion the dog reaches from the punishment is incorrect. For example, say your dog alarm barks when he sees other dogs. You spray him with a squirt bottle, pop his leash, alpha roll, yell, or otherwise punish him when he starts alarm barking. What association may he come to - that alarm barking gets Bad Things for Dogs, or that the presence of other dogs produces Bad Things for Dogs? It's a dangerous game to play as you may create a dog that alarm barks when dogs are further away, ups his display, or eventually has an aggressive response when one comes too close. All because he things the presence of other dogs is now bad.

So, in the end. Punishment based training can work, sort of. It can suppress behaviors, and may make a dog too darned scared to do anything. But we don't have to use punishment to get what we want. We do not have to scare, intimidate, physically dominate, or otherwise cause trauma to, a dog to get it to do what we want. And that's really what matters, it's not that the use of aversives in training may or may not produce results, it's that we do not have to use them.

 I got the full article by requesting it through my college, but I haven't had a chance to read it yet. I'm skimming for better definitions of the techniques.

confrontational methods are described as, "positive punishment, i.e. punishment using an aversive stimulus, such as pain."

Aversive: Direct Confrontation includes

alpha roll

dominance down

force down with leash

hit or kick dog

grab jowls/scruff

knee dog in chest for jumping

neck jab

choke or pronged pinch collar

leash correction

rub dogs nose in house soiled areas

force release of an item in dog's mouth

remote activated shock collar

bark activated shock collar

muzzle

Aversive: indirect confrontation

yell no

quick spray with bottle/pistol

"schttt"

growl at dog

force exposure

verbal punishment for house soiling

"stare down"

Non aversive: reward based training

"look" or "watch me"

clicker training

food rewards

use food to trade for items

 food stuffed toys

sit for everything

Neutral

Avoidance

synthetic pheromones

increase exercise

There are definitions or examples available for most of these if anyone would like me to post those, or if there is any information anyone would like me to find in the article.

 Thanks for clarifying.  I have some major confusion over why some things are in the categories they are in.

 I am not finding anything in the methods section about how the methods where I would think to look for it, about how these were classified. I do see later however what basically suggests that the aversive: direct confrontation appears to involve "physically manipulative techniques."

From reading the whole study, I don't think we're seeing anything that most of us here aren't aware of. If anything, we do have some statistics about the potential harm being caused by techniques recommended on certain shows.
"dogs presenting with aggression to familiar people were more likely than dogs with other presenting complaints to respond aggressively to the alpha roll (100 vs. 50%)"

"there were no statistically significant differences to any of the other interventions between dogs presenting for aggression to people ( both familiar and/or unfamiliar) and dogs with other presenting complaints."

So, if your dog is aggressing towards someone in the family, if you alpha roll, you will get aggression. If it's aggression towards people you don't know, you've got a 50/50 chance of aggression if you alpha roll. However, the highest percentage of aggression came from hitting or kicking the dog, followed by growling at the dog.

Clicker training had no aggressive responses (surprise, surprise). Nor did look/watch me, increase exercise, pheromones.

The part were this study gets dicey though, is in the results. The positive, negative, no results totally do not match with the aggressive responses thing. Something here suggests to me that definitions might not have been excellent, and replicability of this study could be difficult. It appears that if you reported an aggressive response, you may not have been locked in to negative effect. It may have been possible to have had aggression and a positive effect. That seems a bit wrong to me.

tenna

As punishment does not cause a behavior to become extinct, the behavior will eventually pop back up unless an incompatible behavior is trained in the interim.

I have to disagree with you there. I know many dogs who have been trained to break inappropriate behaviours with aversives, why would it pop back up if the dog has learnt that the behaviour is not rewarding?

Also, aversives and punishment cause a dog to associate you with bad things. I don't want my dog to think I produce Bad Things for Dogs. Why would I want him to be afraid of me? Why wouldn't I instead want to create a positive environment where he feels safe and comfortable with me, where he knows I won't (because I never have) yell at him, hurt him, or scare him?

I think you're making quite an assumption there. Why do aversives or punishment have to involve yelling at, hurting or scaring a dog? I've used aversives and punishments with my dogs and they are far from fearful of me (keeping in mind that an aversive can include anything the dog finds remotely unpleasant, such as a vocal correction). Those aversives have never included using pain, fear or intimidation.

Sure, aversives can produce results. They can suppress behaviors you don't like (and ones you do!), but we have better, safer, more humane options out there. Why result to causing a dog discomfort, pain, or fear, when we do not have to?

Again, why do using avesives have to create pain or fear in a dog? Your post makes me suspect you've never learnt how to use them appropriately ;) I'd never use inhumane training methods with my dogs. I'd never use fear to train my dogs. But I'm not ignorant enough to think that I've never used any training method that's not in some way aversive to them.

There is also the issue that the conclusion the dog reaches from the punishment is incorrect. For example, say your dog alarm barks when he sees other dogs. You spray him with a squirt bottle, pop his leash, alpha roll, yell, or otherwise punish him when he starts alarm barking. What association may he come to - that alarm barking gets Bad Things for Dogs, or that the presence of other dogs produces Bad Things for Dogs? It's a dangerous game to play as you may create a dog that alarm barks when dogs are further away, ups his display, or eventually has an aggressive response when one comes too close. All because he things the presence of other dogs is now bad.

You're example is an issue of poor timing, not an issue that arises with the use of aversives and/or punishment when they are used appropriately and correctly.

We do not have to scare, intimidate, physically dominate, or otherwise cause trauma to, a dog to get it to do what we want.

No trainer I've ever known who uses aversives appropriately uses them to scare, intimidate, physically dominate or cause trauma to the dogs they are training.

huski

I have to disagree with you there. I know many dogs who have been trained to break inappropriate behaviours with aversives, why would it pop back up if the dog has learnt that the behaviour is not rewarding?

Because, in reality that's not how it works. Punishment, on its own, still needs punishment (even if not often) in order to keep it suppressed. If the punishment stops, the behaviour will be brought back out and used again, if another, suitable behaviour was not reinforced to replace it. It works the same way as reinforcement - in order to keep behaviours, you have to reinforce them periodically. Punishment is no different - in order to keep them away, you need to use that punishment periodically or the behaviour will crop back up. But, if you've taught a different, useful behaviour to do instead of the punished one, chances are the punished one will go away. It won't truly go extinct though - if you stop reinforcing the desired behaviour, chances are often great the older, once-punished behaviour will come back to the forefront. Punishment doesn't magically "delete" behaviours from a dog's repertoire - they are always there. Dogs, like people, choose which behaviours to exhibit based on their history, and will always have that "original" behaviour to fall back on if circumstances allow.

huski

tenna

As punishment does not cause a behavior to become extinct, the behavior will eventually pop back up unless an incompatible behavior is trained in the interim.

I have to disagree with you there. I know many dogs who have been trained to break inappropriate behaviours with aversives, why would it pop back up if the dog has learnt that the behaviour is not rewarding?

Also, aversives and punishment cause a dog to associate you with bad things. I don't want my dog to think I produce Bad Things for Dogs. Why would I want him to be afraid of me? Why wouldn't I instead want to create a positive environment where he feels safe and comfortable with me, where he knows I won't (because I never have) yell at him, hurt him, or scare him?

I think you're making quite an assumption there. Why do aversives or punishment have to involve yelling at, hurting or scaring a dog? I've used aversives and punishments with my dogs and they are far from fearful of me (keeping in mind that an aversive can include anything the dog finds remotely unpleasant, such as a vocal correction). Those aversives have never included using pain, fear or intimidation.

Sure, aversives can produce results. They can suppress behaviors you don't like (and ones you do!), but we have better, safer, more humane options out there. Why result to causing a dog discomfort, pain, or fear, when we do not have to?

Again, why do using avesives have to create pain or fear in a dog? Your post makes me suspect you've never learnt how to use them appropriately ;) I'd never use inhumane training methods with my dogs. I'd never use fear to train my dogs. But I'm not ignorant enough to think that I've never used any training method that's not in some way aversive to them.

There is also the issue that the conclusion the dog reaches from the punishment is incorrect. For example, say your dog alarm barks when he sees other dogs. You spray him with a squirt bottle, pop his leash, alpha roll, yell, or otherwise punish him when he starts alarm barking. What association may he come to - that alarm barking gets Bad Things for Dogs, or that the presence of other dogs produces Bad Things for Dogs? It's a dangerous game to play as you may create a dog that alarm barks when dogs are further away, ups his display, or eventually has an aggressive response when one comes too close. All because he things the presence of other dogs is now bad.

You're example is an issue of poor timing, not an issue that arises with the use of aversives and/or punishment when they are used appropriately and correctly.

We do not have to scare, intimidate, physically dominate, or otherwise cause trauma to, a dog to get it to do what we want.

 

No trainer I've ever known who uses aversives appropriately uses them to scare, intimidate, physically dominate or cause trauma to the dogs they are training.

What, then, is your definition of aversive?  According to the American Heritage Dictionary of the English Language: Aversion: Causing avoidance of a thing, situation, or behavior by using an unpleasant or punishing stimulus, as in techniques of behavior modification.  "Unpleasant" is not the same as benign, and while I grant that some things are more unpleasant than others to dogs, the very nature of aversion is to intimidate the dog into ceasing the behavior you don't want.  And, please provide an explanation of what you consider appropriate uses of aversion.  It's not enough to tell us that trainers you know are appropriately using it if you won't give us a few examples.  Are e-collars appropriate?  Are prongs appropriate. and, if so, used how?  Get specific.

As to timing, the timing of a punishment STILL cannot prevent the dog from making an erroneous association if the behavior that is being corrected and the appearance of some other stimulus happen simultaneously, which is why leash reactive dogs often get worse, not better, for being corrected when they transgress.  It is because the stimulus (presence of another dog) is still extant when the correction (for barking at the other dog) is issued.  The dog could figure that you don't want him to bark, but he also could just as easily figure that he is punished whenever he is in the presence of an approaching dog (with the result that he begins to really hate other dogs), thus continues to be upset and reactive when they appear.

Dogs that have been trained with aversive methods do not always appear frightened of their handlers.  But, they do tend to develop "learned helplessness" which is one reason they look so obedient, and they are, but try teaching them to put their toys away...most correction trained dogs will sit there waiting to be told what to do, looking very obedient, but not very thrilled to offer any new behavior.  They are not as likely to try to figure it out on their own just to please you.  Rather, they are very concerned with not making any false moves, lest a punishment be issued.  And, we aren't necessarily talking e-collar shocks here.  Dogs learn to inhibit their behavior with simple leash pops, too.  But, take a look at any clicker savvy dog - it's so friggin' easy to teach them new things it would make your head spin!  Just for fun, have a look at some behaviors that aren't your typical boring heeling pattern.  http://www.youtube.com/user/kikopup#p/a/u/1/2Y2DBrqw-O8   Tricks are not useless "less than" fun for people who don't compete in Schutzhund lol - they are endless opportunities to test your ability to teach a dog new or complex behaviors.  Try teaching the behaviors in that video using aversion...pfft.

spiritdogs

As to timing, the timing of a punishment STILL cannot prevent the dog from making an erroneous association if the behavior that is being corrected and the appearance of some other stimulus happen simultaneously, which is why leash reactive dogs often get worse, not better, for being corrected when they transgress.

Common problem but a problem with management of the environment and the person doing the training, not the method itself.  You could say the same about positive reinforcement.  I can't tell you how many dogs I've seen that won't do a down unless it's asked to sit first or will sit as part of the down command.  The handler messed up the training.  I don't think it's any easier to mess up aversive training, it just has worse effects on the dog (IMO, sitting while downing is not as "bad" as amping up your dog on a prong collar in front of another dog).  But, like I keep on saying, no good trainer should tailor their methods based on other people's stupid mistakes.  I still use a clicker to train the sit and the down, but train these behaviors separately and from all positions.