Clicker Training Research

I thought of this thread and everything we've been talking about. When I whip out Koji's harness, he bows his head like I'm going to kill him or something. No clue why as harness should mean fun stuff, not stuff like vet. But for whatever reason, he hates. I always feel bad.

So last night, I finally said no more. I grabbed his harness, the clicker and some treat. At first I was luring him into the harness, basically he needs to put his head through the circle and that's it. And I realized, no no, I'm not supposed to be doing that. So I backtracked even more. I started with Koji acknowledging the harness, sniffing it, nosing it, etc.

Then I started clicking when he'd put more of his nose inside, etc, until he would put his head into it. Then I put him in a sit/stay then released and when released, used my new command "harness up" as he came and put his head through. That took a few times as he'd go around, and I realized he wasn't fully getting it yet. Backstepped a bit to proof that he needs to put his head through, then moved forward again.

Within 10 minutes, my little boy has a new command. I am reinforcing it every time I put his harness on, so hopefully it'll settle in. And since he is now associating harness with food, he's much happier to put his head through it. :)

The point of this post.....even a somewhat experienced trainer (sort of), who's read so much, still does the wrong things. Luring isn't wrong, but I know Koji learns faster with shaping. I also moved forward too fast and had to go back. I didn't break things down enough either. Typical errors for beginners, that I make constantly because I want a fast result. I mean, I literally said, "oh, enough of this, time to train it". LOL. And started training it. If I had taken a few mins to think it through, I wouldn't have had to backtrack at all.
 
k9kreationz said:
I thought of this thread and everything we've been talking about. When I whip out Koji's harness, he bows his head like I'm going to kill him or something. No clue why as harness should mean fun stuff, not stuff like vet. But for whatever reason, he hates. I always feel bad.

So last night, I finally said no more. I grabbed his harness, the clicker and some treat. At first I was luring him into the harness, basically he needs to put his head through the circle and that's it. And I realized, no no, I'm not supposed to be doing that. So I backtracked even more. I started with Koji acknowledging the harness, sniffing it, nosing it, etc.

Then I started clicking when he'd put more of his nose inside, etc, until he would put his head into it. Then I put him in a sit/stay then released and when released, used my new command "harness up" as he came and put his head through. That took a few times as he'd go around, and I realized he wasn't fully getting it yet. Backstepped a bit to proof that he needs to put his head through, then moved forward again.

Within 10 minutes, my little boy has a new command. I am reinforcing it every time I put his harness on, so hopefully it'll settle in. And since he is now associating harness with food, he's much happier to put his head through it. :)

The point of this post.....even a somewhat experienced trainer (sort of), who's read so much, still does the wrong things. Luring isn't wrong, but I know Koji learns faster with shaping. I also moved forward too fast and had to go back. I didn't break things down enough either. Typical errors for beginners, that I make constantly because I want a fast result. I mean, I literally said, "oh, enough of this, time to train it". LOL. And started training it. If I had taken a few mins to think it through, I wouldn't have had to backtrack at all.
Click to expand...
Bravo!!! been there done that :biggrin2:
Leo hates putting his collar on so I offer reward for putting it. I might just try clicker too, it works better for my boys as well.
 
Mom2Melli said:
I think they pick up the body language. The dogs have no difficulty seeing you still in position or motion with a click but not reaching in the pouch for the treat. When they are done with a whole sequence the way you press the clicker is different, you are reaching for a treat, you smile or nod or something, you shift your weight, your pupils probably change. They know. However, shouldn't the dog regardless of whether you say yes, click, treat, or jackpot play shouldn't the dog always be waiting for the next command? So really no matter what you do you are indicating correctness pay attention for the next task until that is a "done"-type cue meaning session is over. I just saw the extra feedback as keeping it tighter and more encouraged for whatever reason the dog may need that (like my insecure shepherd or the dog fo example).
Click to expand...

When I am training a behaviour (using the clicker) then I would be expecting the dog to be driving the action. He is figuring out for himself what behaviour will earn that reward. My job is to indicate which choices are the right ones and reward that choice.

When he has learned the behaviour, I add the cue word so he learns to associate the behaviour with this cue and finally perform the behaviour when he hears this word.

It is only at this point that the dog is waiting for the command and I am in the driving seat controlling the action. By this stage the dog has learned the behaviour and associates it with the cue word and is consistently performing the behaviour in response to the cue. So he has more or less progressed from learning and is now practicing and proofing.

When my dogs are working on the clicker, I am wanting them to use their own initiative to succeed. If I used it reguarly for feedback in a sequence where I am dictating the next move, I think it would make shaping new behaviours very difficult because the meaning/context would alter. It would mean "that was good - I'll tell you what to do next" rather than "right behaviour, can you do it again?" The dogs would learn to wait for some kind of prompt from me rather than offering a behaviour. Possibly this would be less of a problem with luring? I use shaping so much with Epic that it is essential I don't dilute the meaning of the clicker because he has to really want to figure out how to get the click. He's a prospector with the clicker rather than an employed miner.


Mom2Melli said:
In studies, animals treated not every time exhibit far more drive and make more attempts at doing a known behavior to earn the treat. It's like a slot machine. It might not be every time, but it will come eventually. They learn that. I have had sheep, goats, and llamas. I grain trained them. I clap - they come from even the far corner of my 5 acres - there is grain. I reinforce that from time to time just to keep it up. However, there are times when I can clap and there is not grain. They come anyway, mill around, and wander off. That is enough time to get them close up if there might be a predator, to close a gate, or count to make sure there is no escapee. If I never re-treated sure the behavior would stop. However, if they never know if there is grain when I clap, they come every time. Just because they might choose the treat every time doesn't mean if there is sometimes no grain behind X (not behind O but just not there) they would still go to the X the next time hoping.
Click to expand...

This is essentially correct. It does to an extent depend upon the frequency of the reward as well as unpredictability. If the slot machines only rarely paid out people would probably find something more enticing. If grain was only behind X every so often the sheep might give it a try or they might start to thing it wasn't worth the effort.

I would say that your clapping is a cue/signal/command, rather than an indication of sucess like the clicker would be. You clap first and they come in anticipation of grain. They would not come if they didn't hear the clap initially and you would not first give the grain and then clap. The X indicated to the sheep they would find food, it did not tell them they were making the right choice by heading in that particular direction. A gundog whistle is also a signal because the dog is acting as a result of hearing it, rather than trying to make it happen. The slot machine is a signal too - if you play the game you might win. You want to win so you have a go but seeing the slot machine does not guarantee a win. The presence of a reward in response to a
signal doesn't need to be consistant to maintain the behavior. However if there has been an indication of success (flashing lights on the slot machine) then you would know that meant you were right and anticipate a payout. It is different thinking "will I win?" and living in hope to thinking "I've won - where's my money?"
 
Last edited:
Effect on owner

Changing tack a bit here and going onto something which may also be relevant to the OP, has using a clicker changed the way you train or interact with your dogs? Has the change been a positive one or just a different direction?

I have found that using the clicker has meant I don't get impatient, cross or frustrated at all during sessions, which used to happen all the time with traditional methods. Like Toffee'sMom I will go off and rethink the strategy if things aren't progressing, rather than get annoyed with the dog. It's more plesant for the dog but I really enjoy training him too and often miss favourite TV shows because we're really into something. It seems to be a really big motivator from a human perspective - it's FUN and a pleasure rather than a duty.

From an Instructor's point of view I think this is quite important because people are obviously more likely to train their dog if they are enjoying the process rather than finding it a chore out of necessity. And they are more likely to enjoy it if they can see progress happening daily rather than over the course of several weeks or months. It's a big issue because during classes the instructor can only give the owners the information, it is down to them to practice. So many people though just do dog training on class evenings and you can see the idea of homework really doesn't appeal. Frustration with the dog's lack of progress is also a problem - especially because inevitably some dogs do learn quicker than others and people struggling feel their dog is letting them down so the relationship is already under some stress.
 
I am sorry for replying as I really don't want to monopolize this thread any more :no:.

BUT..
I thought that was an interesting article, however I think the extreeme behaviours SG describes are less likely to occur in breeds other than BCs. I think collies are unusual in that they are so ready to obey, they are primed for it so you tell them to jump over that cliff and they pitch out immediatly. Certainly the shelties I've known (mine and others) would at least say "Are you sure about that?" Even the ones that have been falling over themselves to please still have retained their judgement about what is too much. I honestly can't see shelties or other breeds doing the behaviours she describes to that extent, they're too free thinking to be that compliant. So I'm not sure if that is a relevant 'danger' for the majority of clicker trained dogs. Even after 8 years my timing still sucks but nevertheless I have successfully trained a lot of quite advanced behaviours because the dogs seem to get the message as long as it is reasonably clear and they are evidently not as meticulous about exact compliance, which works in my favour. :smile2: Probably explains why I have struggled with micro-shaping though....
 
Just a note: I'm done with the thread so you may wish to stop debating my points repeatedly.

I really don't think a couple extra clicks or yesses during a super short period of time during training warrants that much discussion.

The only thing I want to add -- don't comment on the personality of my dog and suggest that I made him that way. You don't know him and you don't know me. You don't know that he was an orchard puppy who was found in the gutter at 6 weeks alone, starving, full of worms, and was picked up by someone after a car almost hit him and while he was too traumatized to move. You don't know he was in a shelter adotion event pen huddling in a corner afraid of every movement and person coming near so that the adoption rep actually said they didn't know what to do but he was so scared after over a month with them he couldn't get adopted but when he saw me he licked me like I was his long lost soulmate and that his entire family (parents + 3 siblings) were caught later and brought in from the same orchard area so he had been lost or rejected or something but had no human kindness as a pup since the whole family was feral and wild as could be. You don't know that his first time in public he hid under my skirt while my other dog (a therapy dog) proudly sought out people to visit. You don't know that the dog has been through massive amounts of travels, including accompanying my parents and myself to Stanford where my Dad had brain surgery for cancer. You don't know that when I was there a CalTrain flew by and terrified the dog while I was at McDonald's eating outside and that some dumb jerk came up right after that and grabbed him by the collar while I was not looking. The guy did this so he could get close to a little girl who was sitting at the next table alone who I was asked to watch (hence me not looking at the dog) because her grandfather went in to the bathroom. When the dog backed up away from the guy the guy said to the girl "You're pretty, you want to come with me?" at which point the dog and I both about attacked the guy and the grandfather was so grateful he was in tears thanking us for protecting his granddaughter but after that the dog has been afraid of many more things (like sounds of diesels and men) than even he was before despite everything I have done since to build back up his confidence and exposure. While my Quasar loved all people and did Children's Hospital therapy, Holly was a flyball dog with my husband, Mayson did therapy with the elderly in care facilities, and Melli is excelling in obedience classes and improving rapidly by leaps and bounds from an unknown past, don't suggest I did this to Shay. Someone commented that I can't make assumptions based on limited experience somewhere in this thread, well, I am responding the same back. You can differ on training, feeding, housing, grooming, whatever, but don't be making assumptions about MY dog.
 
I've had a busy week so have only now had time to reply, but I'm really sorry Mom2Melli feels she again has to leave the discussion. It seems a great pity to leave this topic in such a negative place. I'm unsure where the percieved critisism of the dog occured - certainly this was never my intention and I apologise if any of my comments have been interpretated in this way. :(
My view was and still is, that the technique is positive and obviously sucessful so not a problem. I just feel the usual process is more effective for the type of communication and training I do. Each to their own. :smile2:

I have personally found this thread very interesting, certainly it has made me think a great deal about the finer points of clicker training and has challenged my preconceptions.

A lot of good points have been raised. In particular it has occured to me that Mom2Melli has hit on something which can be an issue with clicker trained dogs - once the clicker goes, so very often does any kind of feedback. Because a lot of clicker trainers advocate avoiding verbal interaction during the training process (so the 'click' is very clear) people seem to get out of the habit of offering praise/feedback. The dog is expected to perform for longer durations with no support. Although the technicalities have caused some dispute here (!), having given it some thought I don't think the additional clicks (without treats) are as major an issue as I originally believed. That is not to say I would rush out and start doing it, or necessarily advise others to - but it is probably no different in terms of possible confusion than for example poor timing or using the wrong cue, which I'm both guilty of!! :o

If this method works for the handler and dog and they are comfortable with it then that is the most important thing. It is a pity though that the discussion became so defensive. When I started a thread on different agility handling techniques it was interesting to hear how people had adapted or changed from the methods they started out with, even though these were succesful, because they found even better ways of doing it. And the same appears to be true for those that develop the techniques - they may have individual preferences but have been happy to adopt alternatives or make improvements as their system evolved. It is always worth being open minded a new approach (so long as it is positive based training :wink2:) even if initially it goes against long held ideas. It may not ultimatly be the route you want to go down but it's always good to have a think about why something could work or not rather than dismiss it or refuse to change. I think we've all been a bit guilty of this...

One final thing I would like to say is that I went back and re-read Karen Prior's book, Reaching the Animal Mind (Great book!) and lo and behold there is an instance where she too used the marker ( a whistle rather than click), witheld the reward, cued another behaviour which was subsequently whistled and rewarded. She states this reinforced the first behaviour (which she was capturing) with the cue for the second (practicing) and increased the first behaviour while stregthening the second...

While not exactly the same process as Mom2Melli describes, none the less it demonstrated to me that the world wouldn't end if the click wasn't invariably followed with a treat! :winkgrin:
 
Last edited:
One final thing I would like to say is that I went back and re-read Karen Prior's book, Reaching the Animal Mind (Great book!) and lo and behold there is an instance where she too used the marker ( a whistle rather than click), witheld the reward, cued another behaviour which was subsequently whistled and rewarded. She states this reinforced the first behaviour (which she was capturing) with the cue for the second (practicing) and increased the first behaviour while stregthening the second...

While not exactly the same process as Mom2Melli describes, none the less it demonstrated to me that the world wouldn't end if the click wasn't invariably followed with a treat!
Click to expand...

Thank you. The most important thing is what the dog needs to be successful, be it chucking treats at them, clicking once, clicking a couple extra times, screaming in praise with 5 minutes of toy play, kisses, or whatever. Only an owner/handler knows what the dog really needs to perceive praise and strengthen a behavior.
 
You must log in or register to reply here.
Back
Top