In The Wizard of Oz, Dorothy learned that something magical happens when she clicked. In her case, when she clicked the heels of her ruby red slippers together three times, saying “there’s no place like home”, she found herself back in the safety of her Aunty Emm’s house.  Today Sadie (re)learned that something magical happens when she hears a click – she found herself back in the safety of positive reinforcement training.

My guide dog Sadie greets people in the way a dog will naturally greet a human..rushing towards them and jumping up at them to get the attention she craves.  As soon as she gets the attention she wants, she reverts to her normal calm behaviours, those that she has been taught as a pup and during her training to be a working guide dog.  Dogs like humans don’t learn by us wishing it so. We need to teach them what we want them to do and how to behave in ways we find acceptable.    Telling them off and/or correcting unwanted behaviours merely tells them what not to do, not what to do.

The use of a clicker in dog training marks the behaviour that the dog has given that we want from them. They quickly form an association between doing something that we want and hearing a click. Usually this is followed up by a reinforcer – something that the dog finds pleasant and over time will reinforce repeating that behaviour.  In Sadie’s case, any kind of food and any amount of it is a reinforcer. There is a hierarchy in Sadie’s mind of food, so chicken is favoured over a biscuit, a biscuit is favoured over a piece of her kibble etc.  It’s important to recognise that it’s what the dog finds reinforcing that is important, not what a human thinks should be reinforcing to them if you want to use this to teach them.  You can also pair a click with a word so that over time you can fade out the click, but the spoken word elicits the same wanted response. For example, Sadie sits, she hears a click and is given a piece of food.  Next time Sadie sits, hears “sit” along with a click and is given a piece of food.  Repetitions are required to teach the dog clicking is a good thing as it’s followed by their reinforcer, and over time the preferred spoken word with the click being faded out results in the same wanted reinforcer.

When Sadie greets a human, I want her to walk calmly to them, and keep her 4 paws on the floor as she greets them and is patted and spoken to by that person.  I need to teach Sadie this, not least because she has been using her very successful (to her) natural behaviours for 6 years which have got her exactly what she wants (human attention).  It will take time to teach her this and today we made a start on it with the help of a senior trainer from Guide Dogs. This trainer not only trains the dogs to learn the wanted behaviours to safely guide, but also trains the dog trainers to use positive reinforcement methods to teach the dogs their guiding skills.  Guide Dogs have recently moved to only using positive reinforcement so staff and volunteers are still learning these techniques.

It’s a number of years since Sadie was part-trained using a clicker, so the first thing our trainer did today was to test Sadie’s reaction to a click when Sadie sat…without the trainer saying or doing anything to elicit this behaviour from her.  As soon as Sadie sat in front of her (the trainer was sitting on the couch in my living room), she clicked and Sadie got a piece of kibble.  Sadie had got up to receive the kibble, so the trainer waited until Sadie sat back down again and clicked, with Sadie rapidly learning sitting calmly in front of the trainer meant click which meant food. So Sadie (re)learned the magic not only of a click, but the magic to her of positive reinforcement training.  Food!

Next step was to go outside and reinforce using the clicker the behaviours we wanted from Sadie. We start easy to build confidence in the dog, and in our case when Sadie guided me to the kerb and sat down as she has been trained to do (indicating to me that we are at the kerb and a road is immediately in front of us), I clicked and gave Sadie a piece of kibble. Each time she sat at a kerb, I clicked and rewarded the behaviour with kibble. As we progressed on our walk, I clicked and gave Sadie a piece of kibble for any of her learned guiding behaviours progressing to clicking and rewarding when she walked past a human. It’s important to understand from Sadie’s perspective that she needs to learn that calmly walking past (or in time to) a person is the behaviour I want, so we started today with people she doesn’t know and has no interest in to build the association with being calm around a person.

Along with teaching Sadie to be calm in greeting friends and family, I also need to teach her a behaviour that will interrupt her currently preferred method of galloping towards them and jumping up to refocus her on the behaviour I need.  As Sadie was taught a hand touch by her volunteer puppy walker, a vital skill to have if she was to have been matched with a completely blind person who would need her to touch their hand on recall from play in the park before they rewarded her for doing so, we used this.  I’m fortunate in having a small amount of remaining vision given the right environment, but do put my hand out to Sadie if I’m not sure where she is at home or in the garden so I don’t accidentally stand on her or trip over her. Today I matched that with the word “touch“ holding out a closed hand with Sadie touching my hand with her nose, and then being clicked and given a piece of her kibble.  We used this when Sadie’s attention on our walk spontaneously moved from guiding me to something ahead – I didn’t see what it was but my trainer thought she saw a tail disappear under a fence likely belonging to a cat. This enabled me to interrupt Sadie’s attention on the cat and turn it back to me. Over time we will build on this to using it to interrupt any galloping towards a person Sadie knows…well, that’s the theory anyway.  It will take time and perseverance, which for Sadie I have in abundance.

On our way back and nearing home, our trainer went ahead and then reappeared a little way ahead of Sadie.  She has already primed me that when Sadie saw her in the distance if she remained calm and/or turned her attention back to me, I needed to click and give her a piece of her kibble.  Any excited behaviour meant I stood still and said and did nothing,  Because Sadie has learned early on in the walk that keeping her attention on me and giving me the behaviour I needed from her resulted in a click and food, it took her a nano second from recognising our trainer but immediately looking to me because she knew that would elicit food. As we calmly walked towards our trainer, I repeated the click and kibble reward each and every time Sadie looked to me so that we were able to calmly approach our trainer.

Over the next week or so until I meet with our trainer again, I will continue to use the clicker to mark any and all behaviours I want from Sadie.  If we do happen to meet a friend or family member on route before we have had the chance to formally practice calm behaviour in their presence in a training session, I will endeavour to use the new training with Sadie but it won’t do any immediate harm if we don’t succeed in it; it’s highly unlikely that Sadie will be calm in their presence having only spent around an hour being reinforced for being calm in a stranger’s presence in the street. Realistically, Sadie using previously successful means to reach someone she knows in this very early stage of retraining her won’t seriously and irretrievably harm what we learned today.  It’s not possible for a trainer with multiple responsibilities to many dogs and clients to solely focus on one more than once a week.  And, as we are using purely positive reinforcement techniques to train Sadie, I don’t correct unwanted behaviours. Obviously if our immediate safety depended on it, I would find a way to interrupt or stop the dangerous behaviour, but that would in no way derail what Sadie learned today or will learn in the coming days and weeks.

The next stage of our training will be to find a willing family member of friend volunteer that Sadie knows well to stand in a pre-arranged place as Sadie and I walk towards them, slowly decreasing the distance as Sadie gives me my wanted calm behaviour until we reach them.  Her kibble will be swapped for a higher value reinforcer as merely giving her kibble won’t hold her attention on me as her assumed human attention she will get by rushing towards the person is a much higher reinforcer than kibble.

If you’ve read this and wondered why as a human behavioural psychologist I haven’t done this training myself to date with Sadie, to maintain my membership of my professional body I should not do any behaviour modification that I have not been specifically trained to do.  Also, I am highly respectful of the staff in Guide Dogs so will always defer to their skills and experience in training my dogs.