[Curbside Prophet]

http://www.antrozoologisenteret.no/artikler/art_training_methods.pdf

Not that I needed more research to convince me of less problematic methods in dog training, but an interesting read nevertheless.

 

[Marsh Muppet]

In related news: Survey reveals that 96% of social science studies prove what the authors believed before conducting the studies.

 

[Curbside Prophet]

In related news: Survey reveals that 96% of social science studies prove what the authors believed before conducting the studies.

The surveys were completed by pet guardians...these are your authors. So I agree this study reveals what the authors believe.

 

[Poly]

Interesting article.

However, I find a serious deficiency in the report is its lack of consideration of focus training.

Back in the day focus exercises were thought to be important only for dogs that were moving into something else - such as working dogs - and not something to be bothered with for "mere pet training". We now know that it is just the opposite and that focus work is the best foundation for all training.

Dogs that do a lot of specific focus exercises with their handler usually respond well to positive training in all other areas and vice versa. The correlation is one-to-one thus providing a positive feedback for the dogs and for the handler.

Also, dogs that learn and practice focus training generally need much less corrective training as they go along. Their handlers would have much less need to apply corrective actions - although it will be a rare dog indeed that never needs a correction. And whatever corrections are applied would not be seen as "punishment" and therefore not reported as "punishment training".

 

[KBLover]

That was interesting - though I have to say I mix both punishment (I would argue ignoring the dog is punishment, P-, even though they have it under "miscellaneous"...whatever that means...what is a "miscellaneous" method?), and Wally has only 3 of the listed problem behaviors. (fear in few situations - and it used to be many situations, he's improved, barking at dogs and mostly behind fences (he's never barked at a dog not behind a fence), and sometimes he'll eat stuff off the ground.

BTW, why is excitement listed as a problem behavior? Why is it wrong for a dog to be excited?

Heck, if excitement is a problem behavior - then R+ methods cause it in Wally. Especially if we're outside, he'll start getting all wound up and such. Of course, I don't consider that a problem...

 

[Curbside Prophet]

BTW, why is excitement listed as a problem behavior? Why is it wrong for a dog to be excited?

Anytime you talk about emotional responses you have to consider thresholds. Excitement, over threshold, is problematic when that excitement leads to involuntary behavior that is not compatible with what you want. It's also problematic because you can't get to operant behavior (sit at greeting guests, for example) until the respondent behavior (excitement) is addressed. Excitement leads to jumping on guests, which to most of us is not compatible with what we want.

 

[sparkle]

" Traditional techniques have used mainly aversive stimuli, either in the form of positive punishment (application of an aversive stimulus in response to an undesirable behaviour ) or negative reinforcement (removal of an aversive stimulus leading to an increase in the performance of a desirable behaviour) (Lieberman 1999). The use of aversive stumuli in training may have a negative welfare effect implications: it is thought to cause suffering (beerda et al 1997). possibly poses health risks (through increased levels of physiological stress), and has been found to be related to aggression towards other dogs (Roll & Unshelm 1997).


The key word is ***** MAY****** have.

which does not equate to ****WILL**** have.


In my opinion this leaves much for interpretation and variation in the scheme of things when considering individual avenues to gain a particular resolution from a particular approach in various circumstances. There are no absolutes in gaining results when conditioning behavior and this fact does not require a scientific study as does one for a common sense understanding that the use of aversive's ****can*** cause issues. Too often people construct verbiage along with a agenda or ignorance that suggests or indicates that if one uses positive punishment or negative reinforcement such usage ****WILL***** a have negative welfare effect implications that do not justify the means of using such operant’s of conditioning.

I hope other studies are conducted on how dogs might suffer from using any failed approach especially over a extended period of time.

Common sense should remind us and guide us of what nature has to offer in terms of a healthy balance between far left and far right and what happens when we mess with such balances.

 

[KBLover]

Anytime you talk about emotional responses you have to consider thresholds. Excitement, over threshold, is problematic when that excitement leads to involuntary behavior that is not compatible with what you want. It's also problematic because you can't get to operant behavior (sit at greeting guests, for example) until the respondent behavior (excitement) is addressed. Excitement leads to jumping on guests, which to most of us is not compatible with what we want.

Okay, that makes sense - though it seems more correct to say the problem behavior is the jumping, not the excitement, or the call it having too low a threshold (I know most owners won't say that, but those conducting the study could "translate" it), but maybe I'm wrong. If excitement is an emotion, it shouldn't be listed with behaviors, imo. Everything on that list follows this (they are all behaviors or behavior chains).

Besides, you can't really reward or punish an emotion, correct? (Since it's not an operant behavior, I believe you taught me this ) So it seems even more out of place on a study about the results of reward/punishment/"miscellaneous" -based methods. They seem to have made that distinction with the separation-related behaviors being on the list, but not separation anxiety (the emotion) being listed.

" Traditional techniques have used mainly aversive stimuli, either in the form of positive punishment (application of an aversive stimulus in response to an undesirable behaviour ) or negative reinforcement (removal of an aversive stimulus leading to an increase in the performance of a desirable behaviour) (Lieberman 1999). The use of aversive stumuli in training may have a negative welfare effect implications: it is thought to cause suffering (beerda et al 1997). possibly poses health risks (through increased levels of physiological stress), and has been found to be related to aggression towards other dogs (Roll & Unshelm 1997).


The key word is ***** MAY****** have.

which does not equate to ****WILL**** have.

Not to mention that it seems to avoid/ignore a whole leg of the P side of the scale, Negative punishment (i.e. withdrawing something the dog would find/is finding to be rewarding in response to a behavior). This seems only oriented around P+ since it's focused on aversives, but that's only telling half the "punishment story" imo.

 

[Curbside Prophet]

The key word is ***** MAY****** have.

which does not equate to ****WILL**** have.

Too often people construct verbiage along with a agenda or ignorance that suggests or indicates that if one uses positive punishment or negative reinforcement such usage ****WILL***** a have negative welfare effect implications that do not justify the means of using such operant’s of conditioning.

The point of the study is not about the laws of learning theory and its definitions. The laws/definitions do not define aversion.

The point of the study is to highlight that the average dog guardian is ineffective with aversion and aversion adds no long term benefit or is problematic for the average dog guardian. You don't need to read into the study a subconscious message to reason that average dog guardian should avoid aversion...it's virtually intuitive. Unless of course you're speaking of a Utopia where the average dog guardian perfects the use of aversion, but we don't live in that Utopia.

If excitement is an emotion, it shouldn't be listed with behaviors, imo. Everything on that list follows this (they are all behaviors or behavior chains).

The difference is emotion is involuntary behavior. The dog jumps because he's excited. Operant behaviors are voluntary. The dog jumps because his guardian cued "jump". Jumping on guests just happens to be one expression of excitement. Some dogs raise their hackles, bark, pace, or their jaw vibrates...these are all expressions of over-the-top excitement. So are they lumping?...yes. But if they were to list all expressions of excitement it would add more tedium than necessary.

Besides, you can't really reward or punish an emotion, correct? (Since it's not an operant behavior, I believe you taught me this ) So it seems even more out of place on a study about the results of reward/punishment/"miscellaneous" -based methods.

Yes, you can not reinforce emotions. But you can change an association. And the distinction that should be made here is simple...fear based aversion does not lend itself well to classically condition behavior we do want, especially from the average dog guardian.

Not to mention that it seems to avoid/ignore a whole leg of the P side of the scale, Negative punishment (i.e. withdrawing something the dog would find/is finding to be rewarding in response to a behavior). This seems only oriented around P+ since it's focused on aversives, but that's only telling half the "punishment story" imo.

As ineffective as negative punishment is, I'd ignore it too. I don't think the results would demonstrate anything of interest studying negative punishment other than it sometimes works and it sometimes has no effect whatsoever. I believe all the average dog guardian needs to understand is that positive reinforcement works very well, and positive punishment can exacerbate the problem at hand, and offers little (if any) to no long term benefit.

 

[KBLover]

As ineffective as negative punishment is, I'd ignore it too. I don't think the results would demonstrate anything of interest studying negative punishment other than it sometimes works and it sometimes has no effect whatsoever. I believe all the average dog guardian needs to understand is that positive reinforcement works very well, and positive punishment can exacerbate the problem at hand, and offers little (if any) to no long term benefit.

Is negative punishment any more ineffective than positive punishment? I mean, if P+ can actually create problems or make them worse, how much worse off is P- methods?

While I agree the R+ almost always works on dogs, I wouldn't say P- is any more ineffective than P+, except on dog by dog basis. I mean, if I get up and leave for 10 minutes because Wally refuses a command, then I come back and give that command again and get instant and eager compliance - was P- ineffective?

I would say it would be MORE effective than me smacking Wally on the butt or pulling his hair until he yelped, etc, and what negative side effects would me getting up and leaving create - other than it just not working? (And then, if he did it in response to the aversion, wouldn't that be closer R- instead?)

Maybe Wally isn't the "average dog" (what would that be, anyway - just curious?) - but P- is rather effective on him if only to "tell" him he failed to earn the reward.

Of course, P- works when the dog understands the requested behavior - but that would also seem to apply to P+ since no punishment can actually gain behavior.

 

[Curbside Prophet]

Is negative punishment any more ineffective than positive punishment?

I don't know, you'd have to test it in its given application.

I mean, if P+ can actually create problems or make them worse, how much worse off is P- methods?

Lets be clear, physical, scary, painful, harmful P+ makes problems worse. P+ doesn't have to be any of these things, but it is the topic being discussed, however, I don't know any convenient way to use these adjectives to describe P-. So I'd agree P- is not worse than P+ in context.

While I agree the R+ almost always works on dogs, I wouldn't say P- is any more ineffective than P+, except on dog by dog basis.

I don't think the study attempted to illustrate the effective use of learning theory. What the study attempted to illustrate, and perhaps it's too obvious, is that the average dog guardian is not effective with methods x, y, and z. Your average dog guardian couldn't even define P- if they had to.

I mean, if I get up and leave for 10 minutes because Wally refuses a command, then I come back and give that command again and get instant and eager compliance - was P- ineffective?

I'd say it's to be determined. One trial doesn't illustrate anything to me.

what negative side effects would me getting up and leaving create - other than it just not working?

To a velcro dog, perhaps abandonment IS scary to them.

(And then, if he did it in response to the aversion, wouldn't that be closer R- instead?)

The difference between P+ and R- is the target behavior and duration of the aversion. They are always opposite.

"average dog" (what would that be, anyway - just curious?)

To me? Has a tail or once had a tail but that part of the body it is/was attached to wags, has four paws or once had four paws, barks-except if it is a Besenji, has a furred body even on the microscopic level, and likes to sniff bums.

- but P- is rather effective on him if only to "tell" him he failed to earn the reward.

I think if you've had a long, mutually beneficial relationship with a dog, P- can be effective, especially if used in conjunction with a no-reward-marker. Absent of that long, mutually beneficial relationship it's a waste of good training time. My opinion only.

Of course, P- works when the dog understands the requested behavior - but that would also seem to apply to P+ since no punishment can actually gain behavior.

Actually, I'd say P+ works better for known behaviors than P-. Say you want sit (a known behavior). Which is more effective? If I sit!, and the dog sits (P+), or walk out the room for a failed cue, I return, and the dog sits? IMO, the one that saves me time is more effective, but I have to define my criteria for you to understand that point.

 

[harrise]

Ok, I'll ask. ¿How would anyone here go about explaining this to John/Jane Doe dog owner? Arguing in learning theory lexicon is one thing, getting a frustrated wit's end dog owner on board is another.

 

[Curbside Prophet]

Ok, I'll ask. ¿How would anyone here go about explaining this to John/Jane Doe dog owner? Arguing in learning theory lexicon is one thing, getting a frustrated wit's end dog owner on board is another.

Easy, you don't bother with the lexicon. You tell the guardian how you plan on approaching the problem, and ask them if they accept your approach. Many guardians don't care about the methodology, they just want a companionable dog. So the ethics question doesn't reside with the guardian if its your services that you're offering. All you really need is a guarantee from the guardian that they will practice what you teach them. Otherwise, none of it really matters; the guardian's compliance is necessary for effective translation.

 

[KBLover]

Ok, I'll ask. ¿How would anyone here go about explaining this to John/Jane Doe dog owner? Arguing in learning theory lexicon is one thing, getting a frustrated wit's end dog owner on board is another.

Is it really that hard? I think it's just the terms aren't often used, but people use these in all kinds of ways, imo, they don't go "oh I'm P+ you because you didn't get your work in on time!" they just do whatever.

It's not necessary to use the terms, but I don't think it would be all that hard to explain it in "everyday dog owner" language.

I think if you've had a long, mutually beneficial relationship with a dog, P- can be effective, especially if used in conjunction with a no-reward-marker. Absent of that long, mutually beneficial relationship it's a waste of good training time. My opinion only.

That would make sense - the dog would actually have to care that you left him behind/ended interaction with him.

But don't most owners have some kind of good relationship with a dog, other than the utterly clueless and the utterly hostile type owner? Wouldn't most owners have dogs that would actually care if their people stopped interacting with them?

Actually, I'd say P+ works better for known behaviors than P-. Say you want sit (a known behavior). Which is more effective? If I sit!, and the dog sits (P+), or walk out the room for a failed cue, I return, and the dog sits? IMO, the one that saves me time is more effective, but I have to define my criteria for you to understand that point.

Interesting you mention a strongly given command (I'm guessing that's what you were intending - i.e. a harsher/louder/very firm tone of voice), because that used to freeze Wally up and start him to shaking. That's why I went with P- type approaches instead.

Now, I can actually do that - somewhat - but it still shakes him. He gets visably more worried, starts moving slower/looking away (calming signals).

 

[Curbside Prophet]

But don't most owners have some kind of good relationship with a dog, other than the utterly clueless and the utterly hostile type owner? Wouldn't most owners have dogs that would actually care if their people stopped interacting with them?

You forgot the ignorant. I do believe the majority of dog guardians are well meaning people, who seek a mutually beneficial relationship with their dog; but I would also say the majority of dog guardians are absolutely ignorant about what a dog is. And although most dog guardians do want a dog that interacts with them, not all dog guardians appreciate the many number of ways dogs attempt to socialize with their people.

Interesting you mention a strongly given command (I'm guessing that's what you were intending - i.e. a harsher/louder/very firm tone of voice), because that used to freeze Wally up and start him to shaking. That's why I went with P- type approaches instead.

Now, I can actually do that - somewhat - but it still shakes him. He gets visably more worried, starts moving slower/looking away (calming signals).

I have the same dog. I could kick my dog in anger and she'd be looking for a toy to play with afterward. This terrier likes that stuff. But raise my voice and she's emotionally crushed. And guess what? I'm the type of person who naturally insists on things with a raised voice. That's a problem. Using differential reinforcement I had to teach her that my raised voice was really, really rewarding. Along the way I also had to compromise with some patience and learn my sarcastic voice pleased us both. Go figure.

I'm not saying P- can't be effective, I'll use it too to some effect; I just know that I, like many other guardians, want behavior now, not minutes from now. If I cued a sit and my dog didn't sit, my preference is to insist on another sit. If that didn't work, I'd lure a sit, and try to figure out why she didn't sit later. That's what I choose to do. If P- works for you, and you have the time, who am I to tell you any differently?