Variable rates of reinforcement keep people hooked on social media and gambling. When people post on Facebook the little red icon in the top right hand corner that someone has liked our post can be addictive – we sometimes go back and forth to check how many likes this has got. When people gamble, even though they don’t hit the jackpot every time, intermittent jackpots can account for persistent them gambling. On slot machines they create games to keep people hooked and the idea that that you need to keep gambling as it’s fun. Rewards in this form can fuel behaviour and keep people hooked.
Let’s discuss reinforcement and the science of behaviour. In operant conditioning animals learn via association – a cat that meows and immediately gets fed in the kitchen may repeat this next time as they associate their meowing resulting in food. With classical conditioning, BF Skinner conditioned the sound of a bell to result in a dog salivating.
When we are first teaching our dog something new a continuous rate of reinforcement is very important. We positively reinforce a behaviour every time a specific behaviour occurs. For example, every time a dog sits we give them a treat – this helps to instill the behaviour.
With fixed rate reinforcement, behaviour is reinforced only after the behaviour occurs a specified number of times. For example, one reinforcement is given after every so many correct responses, for example after every 5th response. For example, a child receives a sticker for every 3 words they spell correctly.
With fixed interval reinforcement one reinforcement is given after a fixed time interval providing at least one correct response has been made. An example is being paid by the hour.
With variable interval reinforcement if one correct response has been made, reinforcement is given after an unpredictable amount of time has passed, e.g. on average every 5 minutes.
When your first coaching your dog use continual rewards to teach a new behaviour – this will help fix a behaviour. Then use variable rewards once you feel it’s managed to try and stick the behaviour.
Variable reward (Variable Interval Reinforcement) is successful as it uses the dog’s mind and keeps them guessing about when the reward will be delivered in a similar way to people when they go on slot machines as they don’t know when they may next get a jackpot. The slot machines are shiny, colourful and make lots of noise and you have to press a button or pull a lever to take part.
With variable reward scheduling the dog has no idea when he is going to get a treat or win big time with a jackpot reward (hello liver paste!) or different type of reward. This makes his interest larger and he is more motivated. Research has shown that pigeon’s presented with food on a lever that gives unpredictable food rewards will peck the button twice as much as when the reward arrives in a predictable manner.
So once you reliably have taught your dog a behaviour don’t reward them every single time – keep them guessing about when you will next reward them (once you’ve cemented the training). Random rewards can be very powerful.
References
Pippa Mattison, Total Recall
Canine Principles Course