Best Way To Teach Name/operant Conditioning With Greys

[Guest rarmstrong]

So far Willis has learned "come here". I was using it frequently during walks and didn't even mean to teach it. Came to realize with a few trials at home that he know's to come to me when I say it! Such a smarty pants.

 

I'd like to work on his name next. Any suggestions? I have a clicker and treats. I may need to pair the clicker some more. I was using what I thought were good treats (smelly, not too big), but he really wasn't motivated too much by them so it defeated the purpose of the pairing the clicker.

 

Couple things I know from my job as a ABA therapist with kids on the Autism Spectrum,

 

1. Don't pair his name with commands (easy enough, since he only knows come here)

2. In my job, best to incidentally teach it and not force any eye contact or to move myself so that I'm in his range of motion.

 

I understand the theory of operant conditioning. Understand the concepts like the back of my hand, but I'm struggling to apply it to my buddy Willis.

 

Thanks!

[Guest rarmstrong]

So far Willis has learned "come here". I was using it frequently during walks and didn't even mean to teach it. Came to realize with a few trials at home that he know's to come to me when I say it! Such a smarty pants.

 

I'd like to work on his name next. Any suggestions? I have a clicker and treats. I may need to pair the clicker some more. I was using what I thought were good treats (smelly, not too big), but he really wasn't motivated too much by them so it defeated the purpose of the pairing the clicker.

 

Couple things I know from my job as a ABA therapist with kids on the Autism Spectrum,

 

1. Don't pair his name with commands (easy enough, since he only knows come here)

2. In my job, best to incidentally teach it and not force any eye contact or to move myself so that I'm in his range of motion.

 

I understand the theory of operant conditioning. Understand the concepts like the back of my hand, but I'm struggling to apply it to my buddy Willis.

 

Thanks!

[Guest Giselle]

To be honest, you really don't need the clicker if your timing is good and if you're not shaping anything complex. I personally only use a clicker when shaping a very complex behavior that requires a lot of minute rewards.

 

Otherwise, "Name" > pop a treat into his mouth. "Name" > pop a treat into his mouth. etc. etc. etc.

 

I also don't know why yo wouldn't pair his name with commands. Unless he doesn't know any commands at all, I argue that you should pair his name with commands so that you can be sure he is focused on you. This will avoid repetition of commands, which is really what you're trying to avoid.

 

I am not a fan of "capturing" behaviors or incidentally teaching them because the process tends to take too long. I like my dogs to learn something in 5-10 minutes. If they haven't, I'm doing something wrong... And, yes, any and every dog can learn something new in 5-10 minutes A clicker can be handy for this because it tends to excite dogs (they think, "OH boy!! We're going to train! Goodies!!") and tends to build you faster behavioral momentum.

 

Re: Motivation - just like play, dogs can't achieve maximum motivation unless basic needs are met. Is he totally comfortable? If he's stressed in any way, his motivation will not be optimal. This is why we suggest training new dogs somewhere peaceful and quiet, like the bedroom. Is he just full from a recent meal? Try withholding a 1/4 cup of kibble from either breakfast or dinner and use that/other treats to train. Lastly, evaluate your own skills. Are you communicating clearly what you want? What's your criteria? Is it too high/too low? Are you consistently rewarding it? How's your timing? When in doubt, videotape yourself and you'll see clearly what improvements to make. Good luck! Above all, remember that training is supposed to be fun!

Edited February 25, 2014 by Giselle

[Guest rarmstrong]

To be honest, you really don't need the clicker if your timing is good and if you're not shaping anything complex. I personally only use a clicker when shaping a very complex behavior that requires a lot of minute rewards.

 

Otherwise, "Name" > pop a treat into his mouth. "Name" > pop a treat into his mouth. etc. etc. etc.

 

I also don't know why yo wouldn't pair his name with commands. Unless he doesn't know any commands at all, I argue that you should pair his name with commands so that you can be sure he is focused on you. This will avoid repetition of commands, which is really what you're trying to avoid.

 

I am not a fan of "capturing" behaviors or incidentally teaching them because the process tends to take too long. I like my dogs to learn something in 5-10 minutes. If they haven't, I'm doing something wrong... And, yes, any and every dog can learn something new in 5-10 minutes A clicker can be handy for this because it tends to excite dogs (they think, "OH boy!! We're going to train! Goodies!!") and tends to build you faster behavioral momentum.

Hmmm, thanks for the input! I keep trying to apply what I know from my job to training my dog and am finding that it may not be 100% the same.

 

At least in my job, we don't pair names with "work" (discrete trials, table work, etc) because for many of the kiddos I see their name is aversive due to listener responding deficits and using their name to capture their attention to being work starts to pair their name with work. Makes their name aversive, which is the opposite of what we are trying to teach when they lack the skill to make eye contact/respond to their name.

 

Yeaa.....incidental teaching takes FOREVER. And you have to have the perfect storm of lightening fast timing and huge MO to make it salient. Although, having a huge, natural MO is usually the best and most functional way to teach. Kind of a catch 22.

 

I'd like to keep training short and exciting, so I definitely will keep trying with the clicker. Maybe not so much with his name though. I'm about to feed him, so I'm sure this will be a huge opportunity to teach his name

[NeylasMom]

Hmmm, thanks for the input! I keep trying to apply what I know from my job to training my dog and am finding that it may not be 100% the same.

 

At least in my job, we don't pair names with "work" (discrete trials, table work, etc) because for many of the kiddos I see their name is aversive due to listener responding deficits and using their name to capture their attention to being work starts to pair their name with work. Makes their name aversive, which is the opposite of what we are trying to teach when they lack the skill to make eye contact/respond to their name.

If you use reward-based training, your dog doesn't view responding to cues (I much prefer the term cue to command in part for this very reason) as work, they view them as an opportunity to earn rewards. Cues are just words we assign the let the dog know what behavior we'd like them to do. Your dog doesn't care if you use the word "sit" or "blue cheese" to mean put your butt on the ground, they don't understand the English language, they just learn over time that if I do this behavior (put my butt on the ground) when my person says this word ("sit", or "blue cheese", your choice ) then it's very likely I will receive a reward for doing so. Over time, with the continual pairing of the reward with the behavior in response to the cue, the cue actually becomes a reinforcer itself. This is why behavior chaining works.

 

As for your second point, no you don't want to force anything, but I would focus on getting and reinforcing eye contact. Teach name response is as simple as waiting for the dog to make eye contact with you, clicking and treating. Once the dog is giving me prolonged eye contact (ie. he's catching on), I'll start tossing the treat away and onto the ground after clicking. After he eats the treat, his most likely response is going to be to look back at you to see if you've got more treats. Click and treat. When you're fairly positive that he's going ot look back at you after getting the treat, say his name right before he looks back (when he's almost done eating). Now you've got name recall on cue. Then you can move on to building duration.

 

ETA: I like the clicker for teaching name response, at least initially because at the very beginning you may only get a quick glance from him. Your training will go much faster if you can click at the moment he gives you that eye contact. Once he's got the idea the clicker becomes less important. And when you start building duration it probably makes less sense - at that point, verbally praising for continued eye contact with periodic feeding (that you space out over long durations as you progress) is sufficient.

Edited February 25, 2014 by NeylasMom

[PatricksMom]

If you use reward-based training, your dog doesn't view responding to cues (I much prefer the term cue to command in part for this very reason) as work, they view them as an opportunity to earn rewards. Cues are just words we assign the let the dog know what behavior we'd like them to do. Your dog doesn't care if you use the word "sit" or "blue cheese" to mean put your butt on the ground, they don't understand the English language, they just learn over time that if I do this behavior (put my butt on the ground) when my person says this word ("sit", or "blue cheese", your choice ) then it's very likely I will receive a reward for doing so. Over time, with the continual pairing of the reward with the behavior in response to the cue, the cue actually becomes a reinforcer itself. This is why behavior chaining works.

 

As for your second point, no you don't want to force anything, but I would focus on getting and reinforcing eye contact. Teach name response is as simple as waiting for the dog to make eye contact with you, clicking and treating. Once the dog is giving me prolonged eye contact (ie. he's catching on), I'll start tossing the treat away and onto the ground after clicking. After he eats the treat, his most likely response is going to be to look back at you to see if you've got more treats. Click and treat. When you're fairly positive that he's going ot look back at you after getting the treat, say his name right before he looks back (when he's almost done eating). Now you've got name recall on cue. Then you can move on to building duration.

 

ETA: I like the clicker for teaching name response, at least initially because at the very beginning you may only get a quick glance from him. Your training will go much faster if you can click at the moment he gives you that eye contact. Once he's got the idea the clicker becomes less important. And when you start building duration it probably makes less sense - at that point, verbally praising for continued eye contact with periodic feeding (that you space out over long durations as you progress) is sufficient.

If I can intrude, I thought that dogs found direct eye contact to be threatening? We've been training Leo and Henry when we do"look at me" to look at our hand on our chest, not to look us right in the eyes. Is this incorrect?

Edited February 25, 2014 by PatricksMom

[Guest rarmstrong]

If you use reward-based training, your dog doesn't view responding to cues (I much prefer the term cue to command in part for this very reason) as work, they view them as an opportunity to earn rewards. Cues are just words we assign the let the dog know what behavior we'd like them to do. Your dog doesn't care if you use the word "sit" or "blue cheese" to mean put your butt on the ground, they don't understand the English language, they just learn over time that if I do this behavior (put my butt on the ground) when my person says this word ("sit", or "blue cheese", your choice ) then it's very likely I will receive a reward for doing so. Over time, with the continual pairing of the reward with the behavior in response to the cue, the cue actually becomes a reinforcer itself. This is why behavior chaining works.

 

As for your second point, no you don't want to force anything, but I would focus on getting and reinforcing eye contact. Teach name response is as simple as waiting for the dog to make eye contact with you, clicking and treating. Once the dog is giving me prolonged eye contact (ie. he's catching on), I'll start tossing the treat away and onto the ground after clicking. After he eats the treat, his most likely response is going to be to look back at you to see if you've got more treats. Click and treat. When you're fairly positive that he's going ot look back at you after getting the treat, say his name right before he looks back (when he's almost done eating). Now you've got name recall on cue. Then you can move on to building duration.

 

ETA: I like the clicker for teaching name response, at least initially because at the very beginning you may only get a quick glance from him. Your training will go much faster if you can click at the moment he gives you that eye contact. Once he's got the idea the clicker becomes less important. And when you start building duration it probably makes less sense - at that point, verbally praising for continued eye contact with periodic feeding (that you space out over long durations as you progress) is sufficient.

With ABA, we primarily utilize positive reinforcement. We aim to increase the likelihood that the child will engaged in the desired bx by providing a reinforcer, whether that reinforcer is their favorite snack, video, social praise, etc. Even with positive reinforcement, if the motivation to escape or engage in some other kind of behavior is higher, then there is less likely of a change that they will engage in the appropriate bx. I have kids were you can quickly create a chain in behavior by utilizing their name to capture their attention if an SD is presented shortly after. Regardless of the presentation of a highly preferred reinforcer.

 

Perhaps I'm misunderstanding reward based training and positive reinforcement? Or maybe it's not as complex? Murr

With ABA, we primarily utilize positive reinforcement. We aim to increase the likelihood that the child will engaged in the desired bx by providing a reinforcer, whether that reinforcer is their favorite snack, video, social praise, etc. Even with positive reinforcement, if the motivation to escape or engage in some other kind of behavior is higher, then there is less likely of a change that they will engage in the appropriate bx. I have kids were you can quickly create a chain in behavior by utilizing their name to capture their attention if an SD is presented shortly after. Regardless of the presentation of a highly preferred reinforcer.

 

Perhaps I'm misunderstanding reward based training and positive reinforcement? Or maybe it's not as complex? Murr

And by complex, I mean the bx I'm used to encountering (generally problem bx- automatically reinforcing repetitive bx, escape maintained bx, etc) isn't something I need to be as worried about.

 

I just want to do it the right way :/

Edited February 25, 2014 by rarmstrong

[Guest Giselle]

While behavior modification with children and animals is very similar, children are much more complex and prone to internal motivations, high-processing of thoughts, advanced cognition that we aren't sure occurs in animals, etc. It's similar but not same. "Reward-based training" is just the new popular term for people who primarily use positive reinforcement and negative punishment to train behaviors. Do what I like = you get an incredible reward. If you do something I don't like = the reward for that behavior gets removed.

 

Eye contact: Of course, staring any creature -human or dog- in an offensive, strange, threatening way is NOT what we want. But you can bet your stars that "Eye contact/Focus" is the very first behavior I teach every dog. It is the pinnacle of obedience. You cannot compete in obedience if your dog isn't giving 110% eye contact. It is not only a good thing. It's an incredible and highly desirable behavior. In my experience, people who claim it is "threatening" also tend not to train their dogs at high levels, so take it for what you will. Edit to explain: The reason eye contact is so valuable is because you simply cannot go forth in ANY exercise - agility or obedience or whatever - unless you know/i] you have your dog's undivided attention. Eye contact/Focus are surefire behaviors to ask of a dog to ensure that you have their undivided attention.

Edited February 25, 2014 by Giselle

[Guest Giselle]

 

 

I have kids were you can quickly create a chain in behavior by utilizing their name to capture their attention if an SD is presented shortly after. Regardless of the presentation of a highly preferred reinforcer.

Can you clarify this? From what I'm reading, it sounds like you're describing Premack's principle... which also works in dogs. See: "Come and Go" training.

[Guest rarmstrong]

Can you clarify this? From what I'm reading, it sounds like you're describing Premack's principle... which also works in dogs. See: "Come and Go" training.

Eh, not quite. But that's something we definitely do utilize! We generally use it to introduce play activities that our kiddos don't inherently/aren't internally motivated to play with or play with functionally. For example, I would set it up like this,

-Table work (Less preferred activity)

-New toy

-Highly preferred activity

 

Over time, you phase out the duration of the highly preferred activity as the middle activity becomes more reinforcing. Eventually the activity is reinforcing itself and may move over to a highly preferred activity.

 

What I meant by creating a chain, I more so meant a chain for problem bx. So, if you use the kids name to capture their attention and then present an SD that is perceived as work (varies on the kid and their particular deficit. Its generally discrete trials where you run into this problem-usually a mixture of verbal behavior trials, listener respoding, intraverbal, tacts etc. These skills are really hard to teach in the natural environment so you are more than likely going to have to do it at the table as brief trials) , you can create a chain where the kid engages in problem bx after hearing their name. Their name becomes an antecedent for work, and then you get a funky behavior chain like this

 

-Name, "Sally"--> Problem bx, head banging, vocal stereotopy, etc

 

Thank you though for clarifying and putting into perspective what I should expect when utilizing positive reinforcement training with Willis. I know ABA and autism really well, but it's kind of making me nervous to do things that may seem wrong in my job but aren't wrong when applied to dog training.

[NeylasMom]

To the OP, perhaps you should pick up a basic, solid book on understanding dog behavior/training. I think trying to think of training your dog in context of the behavior modification you do with children (especially children with difficulties) is really blurring your understanding and over complicating your thinking. We're talking about dogs. Yes, operant conditioning is operant conditioning, but dogs are not people. And teaching a dog a basic cue like name recognition isn't behavior modification. A Patricia McConnell book might be a good choice to start looking at things from the perspective of teaching/training a dog.

 

PatricksMom, in terms of dog communication, prolonged intense eye contact (ie. the infamous "hard stare") is seen as confrontational, but we're not dogs and dogs know that! Most dogs learn to find eye contact with humans rewarding, especially when it's paired with reinforcers like really yummy treats. And focus from your dog is really one of the basic building blocks of other training so I do think you should switch to working on getting that eye contact.

[Guest rarmstrong]

To the OP, perhaps you should pick up a basic, solid book on understanding dog behavior/training. I think trying to think of training your dog in context of the behavior modification you do with children (especially children with difficulties) is really blurring your understanding and over complicating your thinking. We're talking about dogs. Yes, operant conditioning is operant conditioning, but dogs are not people. And teaching a dog a basic cue like name recognition isn't behavior modification. A Patricia McConnell book might be a good choice to start looking at things from the perspective of teaching/training a dog.

 

PatricksMom, in terms of dog communication, prolonged intense eye contact (ie. the infamous "hard stare") is seen as confrontational, but we're not dogs and dogs know that! Most dogs learn to find eye contact with humans rewarding, especially when it's paired with reinforcers like really yummy treats. And focus from your dog is really one of the basic building blocks of other training so I do think you should switch to working on getting that eye contact.

Thanks for the suggestion. It's hard for me to get out of work mode sometimes :/ I tend to see everything that way...poor boyfriend haha

 

I have The Power of Positive Dog Training by Pat Miller. Would you suggest that book? I'm working my way through it.

[NeylasMom]

I love Pat Miller, have trained with her. It's a great book, as are all of her books. But its a little more skill focused. I would add a McConnell book to your repertoire for more of the understanding the dog-human relationship kind of stuff. The Other End of the Leash is her most popular, but For The Love of a Dog is also a good read.

[PatricksMom]

Giselle, Neylasmom thank you for the explanation.

[NeylasMom]

Giselle, Neylasmom thank you for the explanation.

I did mean to add that if I felt threatened by a dog in some way, I would definitely avoid eye contact. Same is true of a fearful dog who I am trying to get comfortable with me. In those cases I would try to mimic non-threatening dog body language as much as possible. But for a pet dog who you have a bond with and are just doing some basic training with, its a good thing to reinforce that eye contact is positive through name response or look at me. Just wanted to give that caveat lest someone go out and get bitten by a stray dog.

[Guest rarmstrong]

Thanks again for the input!

 

I enrolled myself and Willis for an obedience class. We start on Sunday. Hopefully that, plus Pat Miller's book and Patricia McConnell's book (which I got as an ebook ) we will bond better. I will learn how to read him better, and the best way to teach/relate to him.

 

I love to learn, and I LOVE learning about behavior (I have an interview for the Clinical Behavioral MS in psychology program at Eastern Michigan University in a couple weeks. They are big on teaching behavioral therapy, and offer specific classes and practicums so you can earn your BCBA) so I'm super excited!

Edited February 25, 2014 by rarmstrong

[jetcitywoman]

I just want to add a couple thoughts:

 

Re: name training

I think this is the easiest thing in the world to teach a dog and IMO the first thing one should teach a new dog. You start out as Giselle described: name/treat, name/treat. After a few repetitions you can add a short delay between name and treat. A little longer again and you'll notice that if the dog is looking away when you say his name, he will look at you so treat that. Then you deliberately wait for him to look away, say his name and treat him when he looks at you. From that you can move on to recall training or other things if you like to use name/command combos.

 

Re: looking in the eye

I'm not a dog trainer but just know my relationship and bonding "path" with my dogs, so this is my opinion, but eye contact is very different with a strange dog versus a dog you have a bond with. A dog that loves you will gaze into your eyes with a prolonged stare. I don't know what it means yet, but I do feel it's some method of communication. So when you do "look at me" training whether you target your eyes or your chest may depend on how bonded you already are with the dog.

[Guest Giselle]

I actually do think it is all fundamentally the same. However, these are different situations: The difference between asking children to "work" and dogs to train is that, when done correctly, dogs LOVE to train. Children may never really truly enjoy "work". I study all the time because it is a highly trained behavior, but it's definitely not something I intrinsically enjoy. I can make studying secondarily rewarding because I know it leads to good grades, but studying itself is an aversive behavior to perform. VERSUS dogs love to train because it is fun, it produces food, and it is engaging. It is extrinsically and probably intrinsically motivating. It sounds like your kid-students have learned to make negative associations with work, and so the undesirable behavior chain has been created. In the end, it is the same. If these principles weren't all the same, we wouldn't call this a "science"!! Just be careful of how you're thinking about these things.

 

For the nitty gritty on dog cognition, I recommend work by Alexandra Horowitz or Ray Coppinger. For easier to read popular stuff, Patricia McConnell is a good introduction. I think you would benefit most from Bailey/Burch's "How Dogs Think", though, since you already have a behavior background.

[NeylasMom]

The Bailey/Burch book is How Dogs Learn. There is a Stanley Coren book called How Dogs Think. Both are great. I almost suggested How Dogs Learn, but thought it might be familiar info already for the OP, not sure.

Edited February 25, 2014 by NeylasMom

[GeorgeofNE]

My two cents: the command should be "come." Not "come here." Shorter is better.

[greysmom]

Couple thoughts. Not that I can add much to the professional part!

 

Working with "normal" dogs is fairly easy - name+treat often enough and he'll learn his name. But I have what's called a "spook." She has what is called "greyhound idiopathic fear syndrome" which is a fancy way of saying that life scares her. She has a chemical inbalance in her brain that makes her highly anxious in everyday situations. She's gotten much better over time, and with medication. My point though is that working with her is very much like working with autistic children.

 

A friend of mine has an autistic child and one day as he was talking about his son's behaviors, it came to me that Cash is very much like this - she doesn't like to make eye contact, she often doesn't respond to her name, she doesn't like to be touched, she has obsessive behaviors that she repeats over and over, she often seems like she is in her own world, food is an big issue for her. I have had success in using conditioning with her such as you and my friend describe for autistic children rather than normal dog training.

 

One of the first things I had success teaching her was a "watch me" cue. She spent months hiding in our bedroom, shaking like a leaf, not looking at anyone. But the day she finally figured out that looking me in the eye got her a reward, she began to blossom. She gained confidence and began to relax and get settled in - finally! When we discovered the right combination of medication to help her, she really came out of her shell.

 

So I really used both the conditioning and training cues you were discussing!