There are nearly as many training methods as there are dog trainers, however all dog trainers use the same basic principles of training whether they are aware of it or not. Dogs that are “positive only” trained generally use either operant conditioning (using food as a reward when the dog correctly performs a task) or classical conditioning if the trainer uses a clicker or mark (pairing the reward with the mark or clicker) Old school dog trainers use positive punishment and negative reinforcement. Positive punishment works by presenting a negative consequence (collar jerk, electric shock, neck pinch) when an undesired behavior is exhibited, making the behavior less likely to happen in the future. When the dog presents the correct behavior the unpleasant stimulus stops (negative reinforcement). This reinforces the correct behaviour by removing the unpleasant stimulus (such a stopping jerking on the collar/electric shop, pinch).
Most dog trainers use a combination of the above methods. They are all well proven to work and there are many papers written on the effectiveness of each (which I will cover in a future blog post). Which method you use depends on a few things, including your personal experiences, who you train with, what training knowledge you have and the dog/s you are training.
Many sheepdog trainers use another method which is rarely used in general dog training: the model-rival method. In this method dogs are allowed to watch another dog perform a task and are then allowed to attempt the task. This method was very successfully used in my search and rescue training and helped in lifting the dogs drive and enthusiasm for tasks.
Allowing your dog to watch other dogs perform and be rewarded for completing training exercises is definitely beneficial and has great potential in teaching a dog tasks quickly and building enthusiasm for that task. When training your dog you can always hold, tether or crate your dog in a location where it can observe other dogs working. Of course different dogs learn at different rates and for some dogs this method would not be as efficient as classical training. Many trainers would also be embarrassed to leave their dogs in such a position as their dogs will become very simulated and bark excessively (being careful that this barking is in fact enthusiasm and not anxiety). For the average pet dog handler this might be considered bad behaviour but for many trainers out there this excited behaviour means that you have a dog that is motivated and ready to learn.
The attached paper took this training one step further. They set up an exercise where two people sat opposite each other with the dog tethered close by. The people then handed an object (sock) back and forth to each other in an animated way. At each pass the person would say “would you like this SOCK?” and the other would say “thankyou for the SOCK”. After two minutes the dog would be asked to retrieve the “sock”. The researchers found that the dog had learnt the name of the object through pure observation and would be able to retrieve the object by name from a group of 3 similar objects. I assume that all objects had the same handlers scent on them to remove scent discrimination. The researchers found that the model-rival method works as fast and accurately at training a dog a task as operant conditioning. The beauty of this training method was that is used intrinsic reward (the object or task is the reward) as opposed to operant conditioning which uses extrinsic rewards (food/toy is the reward).
There are so many applications for this type of training. In detection it could be easily used when training two dogs by allowing one to observe the other in a search. This type of training will also help a younger dog who might be environmentally sensitive to be less so or erratic in their search patterns to settle quicker. A young dog also may learn search patterns/tactic form observation of an older/more experienced/settled dog.
Most dog trainers use a combination of the above methods. They are all well proven to work and there are many papers written on the effectiveness of each (which I will cover in a future blog post). Which method you use depends on a few things, including your personal experiences, who you train with, what training knowledge you have and the dog/s you are training.
Many sheepdog trainers use another method which is rarely used in general dog training: the model-rival method. In this method dogs are allowed to watch another dog perform a task and are then allowed to attempt the task. This method was very successfully used in my search and rescue training and helped in lifting the dogs drive and enthusiasm for tasks.
Allowing your dog to watch other dogs perform and be rewarded for completing training exercises is definitely beneficial and has great potential in teaching a dog tasks quickly and building enthusiasm for that task. When training your dog you can always hold, tether or crate your dog in a location where it can observe other dogs working. Of course different dogs learn at different rates and for some dogs this method would not be as efficient as classical training. Many trainers would also be embarrassed to leave their dogs in such a position as their dogs will become very simulated and bark excessively (being careful that this barking is in fact enthusiasm and not anxiety). For the average pet dog handler this might be considered bad behaviour but for many trainers out there this excited behaviour means that you have a dog that is motivated and ready to learn.
The attached paper took this training one step further. They set up an exercise where two people sat opposite each other with the dog tethered close by. The people then handed an object (sock) back and forth to each other in an animated way. At each pass the person would say “would you like this SOCK?” and the other would say “thankyou for the SOCK”. After two minutes the dog would be asked to retrieve the “sock”. The researchers found that the dog had learnt the name of the object through pure observation and would be able to retrieve the object by name from a group of 3 similar objects. I assume that all objects had the same handlers scent on them to remove scent discrimination. The researchers found that the model-rival method works as fast and accurately at training a dog a task as operant conditioning. The beauty of this training method was that is used intrinsic reward (the object or task is the reward) as opposed to operant conditioning which uses extrinsic rewards (food/toy is the reward).
There are so many applications for this type of training. In detection it could be easily used when training two dogs by allowing one to observe the other in a search. This type of training will also help a younger dog who might be environmentally sensitive to be less so or erratic in their search patterns to settle quicker. A young dog also may learn search patterns/tactic form observation of an older/more experienced/settled dog.