The click is not the trick

[skinnydipper]

The click is not the trick: the efficacy of clickers and other reinforcement methods in training naïve dogs to perform new tasks

https://peerj.com/articles/10881/?f...dhbp3Icvlemh_u_IEZzaUMVSdUuH6GSR3osjVWE4N2D-I

 

[CorvusCorax]

Am I being fick? I don't use a clicker/verbal (yip, good, whatever) as a 'reinforcer', I use it as an a 'mark' that the dog has performed the behaviour I want 'yes, that's correct) and the 'reinforcer' may follow...'reinforcer' for me is the food or the ball or release from the behaviour.

Instantly delivering a reward without a mark/indication that a reward is coming, teaches the dog to react to body language rather than verbals (movement/putting hand in pocket) and I personally don't want that.

 

[ownedbyaconnie]

Am I being fick? I don't use a clicker/verbal (yip, good, whatever) as a 'reinforcer', I use it as an a 'mark' that the dog has performed the behaviour I want 'yes, that's correct) and the 'reinforcer' may follow...'reinforcer' for me is the food or the ball or release from the behaviour.

Instantly delivering a reward without a mark/indication that a reward is coming, teaches the dog to react to body language rather than verbals (movement/putting hand in pocket) and I personally don't want that.

Same. Click and treat is how I was taught. The treat being either praise, scratch, ball or actual food treat.

 

[CorvusCorax]

I'm maybe getting caught up in linguistics. Click can be first reinforcer, reward can be second, etc etc etc
But I've always been taught there's a 'mark' and then a 'reinforcer'. As mentioned, just delivering reward with no indication that it's coming just makes the dog scan you for movement/break the behaviour.

 

[Ambers Echo]

Well of course the click is not the trick. The click simply marks the fact that the dog has performed correctly and will soon get a reward. There are many situations in which you cannot instantly directly reinforce a behaviour so the click allows for that. What a weird paper.

 

[blackcob]

I admit I haven't had time to read it properly yet but there was a bit:

Although no study has directly tested if the clicker functions as a marking stimulus, bridging stimulus, or secondary reinforcer, by definition alone it does not seem as though the clicker could have been functioning as either a marking or bridging stimulus; instead, the clicker appears to function most similar to a secondary reinforcer because it is deployed immediately following the completion of the desired response and is paired to reliably predict the arrival of the primary reinforcer.

I am still none the wiser TBH. That's a bridging stimulus innit?!

 

[Ambers Echo]

Yes as far as I can see. Strictly speaking a 'marking' stimulus is not itself reinforcing - it is just information to the animal. Whereas a secondary reinforcer has become itself reinforcing by being paired with a primary reinforcer. But that really is just semantics because marking stimuli become secondary reinforcers by being paired with a primary reinforcer, whereas those secondary reinforcers stop being reinforcing if you break the link with the primary reinforcer.

So basically I don't get it! They seem to be the same thing in all practical ways.

 

[CorvusCorax]

You're also relying on the skills, deployment and timing of the operators taking part in the study.

The vast majority of handlers move before they mark. They don't even realise they have their hand in their pocket several seconds **before** they've marked or until they've had 'MARK THEN MOVE' or 'I'M GOING TO TIE YOUR HANDS TO YOUR SIDES' roared at them AGAIN.

Er...so I'm told....

But, anyhoo, this is is how movement becomes the primary marker and why a lot of dogs break behaviour when people move their hands.

 

[CorvusCorax]

Who's for Skinner boxes ?

 

[PapaverFollis]

My understanding is that the click becomes a secondary reinforcer. I thought a bridging stimulus was more a "gooooooood, keep going", whereas a click can end the behaviour.

 

[blackcob]

That would make more sense PF, I think we're possibly getting caught up too much in the semantics of it.

The vast majority of handlers move before they mark. They don't even realise they have their hand in their pocket several seconds **before** they've marked or until they've had 'MARK THEN MOVE' or 'I'M GOING TO TIE YOUR HANDS TO YOUR SIDES' roared at them AGAIN.

Can I borrow you to come and smack my hands every time I do this for as long as it takes to stick?

I still have my first 'proper' agility instructor's voice in my head and she's been dead a few years now:

"What's your release cue?"

"Ok."

"Just ok?"

"Yup."

"So why did you also raise and drop your hand?"

"I, er, didn't know I had."

"Well, unless you would like to make flapping your arm an intrinsic part of your dog's start line behaviour, and are going to flap it with consistency, let's try again without shall we."

?

 

[CorvusCorax]

I literally had to put a lead around someone's wrist to stop them lifting their arm up to click ?

 

[deb_l222]

You're also relying on the skills, deployment and timing of the operators taking part in the study.

The vast majority of handlers move before they mark. They don't even realise they have their hand in their pocket several seconds **before** they've marked or until they've had 'MARK THEN MOVE' or 'I'M GOING TO TIE YOUR HANDS TO YOUR SIDES' roared at them AGAIN.

Er...so I'm told....

But, anyhoo, this is is how movement becomes the primary marker and why a lot of dogs break behaviour when people move their hands.

This is exactly the reason I have never attempted clicker training - I'm too twitchy!! Done correctly though, it's really good to watch and definitely has a place in the world of training any animal.

 

[FinnishLapphund]

The biggest thing Clicker training taught me is that regardless which method you use when training a dog, timing is what really matters. You can click as much like, but if your timing is off, you're not rewarding what you want to reward. Which is the same as in any other training. But clickers seems good at helping some owners to become more aware about their timing.

I'm not one of them. I've tried it, but clickers isn't for me, because I forget to use it. Too often it became "Oh no, wait, I forgot to click", when I yet again just said "Good", and gave the bitch a treat/something else.
I'm actually better at using a Shaping whistle (like in dolphin training), that I can hold in one corner of my mouth, rather than a clicker.

Anyhow, even though I occasionally got a bit confused at what they actually meant in some parts of Skinnydipper's link, I thought some of it was quite interesting.
For example:
"They found that clicker-trained piglets acquired the fetch behavior in significantly fewer trials than the verbally-reinforced piglets, but that verbally-reinforced piglets made significantly more correct choices on the discrimination task."
And:
"Additionally, it is possible that the frequency and duration of training sessions may impact behavioral acquisition. Demant et al. (2011) found that dogs trained to sit and stay in a basket once or twice weekly reached a significantly higher acquisition level than those trained daily, as did those trained with only one session as opposed to three consecutive sessions per day."

 

[CorvusCorax]

For me the clicker is a bit of a red herring. As long as there's a sound that means 'yes, that's what I want'.

 

[PapaverFollis]

Hahaha, the final paragraph you quoted, FL, is music to a lazy dog trainer's ears. One session, twice a week. Yeah I can manage that. Hurrah for latent learning.