As I start to coach trainers more and more, one of the first things I talk about is variation within your reinforcement. The idea is that the animals do not know what reinforcement they will get, but they know it will be good. As trainers we focus very much on food as reinforcement because this seemed to work best. I often see people asking on Facebook or animal training groups ‘What favourite reinforcers do people use for an animal ‘. While it might help people get started, to be honest, I don’t believe this will work in the long run. Reinforcement should be varied and should not be depending on favourites.
Animals can get complacent, which could manifest through stereotypic behaviour or other problem behaviours. Humans are very good at being predictable. This gives us a safe foundation to stand on, but for motivation we actually need variety. If we would eat the exact same thing every single day, after a couple days if not the same day, we will get bored of it and want something else. The outcome at the start will be different as the outcome at the end. I’ve worked with animals that when I reinforce the exact same food source all the time, I see a decrease in behaviour. Applying some variety in the reinforcement, increased the behaviour.
I also feel strongly about variability within the training sessions we do with our animals. Does the variability really work? When working in the Netherlands we implemented a variety of reinforcement to our training sessions. The animals responded very well to the changes of the reinforcement by showing more motivation for the sessions, we saw a definite increase in behavioural learning. We are able to add variety in many different ways within the training sessions, from different bridge moments, to various toys/food items, giving them more space and so on.
There is one part I want to focus on. We all know we can motivate animals by food reinforcers, although what if the animal is not hungry? Are we not going to interact or train the animal at all?
What about if we look into what the animal exactly wants? There was a study done with horses to see their preferences of wearing a horse blanket when going outside. FIND OUT MORE HERE. The animals could choose the blanket depending on the weather outside, they were able to make their own decisions. Within our training sessions animals make a ton of decisions on their own. I think we as modern zookeepers, have to open up our minds about what would be more reinforcing for the animals we work with.
The other day I was at the Dolphinarium in Kolmårdens Wildlife Park. I had to retrain a spin-bow with one of our dolphins. I asked 2 approximations and she did the criteria I was looking for. My reinforcement in completing the behaviour was a jackpot (bucket of fish). By giving a jackpot (my entire food supply) most people would think to end the session, I was out of reinforcement or was I? This animal loves playing in the water with the trainer and interacting in a very playful way. This interaction is more reinforcing than the bucket of fish I provided.
People might argue that food is the best reinforcement there is, but personally I disagree with this statement. I believe that food is the most important one, but not the best one. The best one is the variety and the curiosity you give the animal or even the control as a consequence of behaviour. In the video above, you see myself in a training session with one of the dolphins at Kolmårdens Wildlife Park. The idea is that I’m very unpredictable with the consequences I give. I can play many different ways with a ball. I can give many different rubdowns to the animal and I can use my energy to reflect to the animal what we will be doing next. When I stay unpredictable with my signals and precursors of my signals, I can keep training for a while.
Animals give us signals all the time what they enjoy, if this a rubdown, being with another individual or preforming their favourite behaviour. Of course we should know our species very well, but we should look more into what the animal actually wants. Instead of thinking this is what they need to survive, so they will see this as having the best, highest reinforcing value. If we find more reinforcing techniques to use besides food, the variety provided to that animal is bigger and more effective, which will have a direct effect on motivation and the relationships we create with our animals.
The Premack Principle is a technique used that was originally identified by David Premack in 1965.
Premack Principle: “The principle that high-probability behavior reinforces low- probability behavior”. (IMATA Glossery, 2004)
The Premack Principle states – “That high-probability behaviour reinforces low-probability behaviour”. (IMATA 2004) We can also explain it as, “more probable behaviours, reinforce less probable behaviours” .
In other cases, through proper observation we can discover what our animals would like exactly and use this as reinforcers. When we observe an animal wanting to be with another individual we can ask a more challenging behaviour and reinforce this behaviour by giving the other individual as reinforcement.
The Premack principle has advanced and enriched the thinking about reinforcement in significant ways. With The Premack principle any behaviour could serve as a reinforcer, provided that it is more likely than the previous performed. For higher levels of motivation to perform a particular behaviour, we can give a more desirable activity as a consequence. This will change the motivation of the animal completely, especially if we are spot on with what the animal exactly wants in that particular moment.
Getting to know your animals is one thing, understanding what they want, is another.
Variability is the key. This can be achieved with different types of food or we can step it up and start looking at what the animal desires at that moment. The balance of reinforcement and variability will give you animals that have more focus and want to learn with you.