"The purpose of variable reinforcement is two fold:
1) increase variability (and consequently improve the attention of the animals) and 2) provide an increased margin of safety for trainers and keepers by eliminating the negative (sometimes aggressive) reaction from animals on continuous reinforcement when, for some reason, they are not reinforced.
In all situations, correctly performed behaviors are bridged. The variability occurs because of whether or not the trainer delivers the primary reinforcer (usually food). If you can reach the point when you are reinforcing only one out of two, or two out of three correct behaviors, you've done your job. There is really no reason to push it beyond that point. Here are some tips on how to implement variable reinforcement with a minimum of stress for you and the animals.
1. Try to avoid 'extinction burst' (the frustration driven, burst of energy caused by non-reinforcement of a correctly performed behavior). This is a particularly common phenomenon with animals on a continuous schedule of reinforcement. A good way to start is to have the animal do a behavior of medium difficulty bridge - and then quickly give the Sd for a fast, easy behavior (a head shake or vocal, for example). If it is performed correctly, bridge again and immediately reinforce. This lessens the amount of time the animal must wait for reinforcement after a non-reinforced behavior and reduces the chance of a negative reaction. (The obvious corollary is that bridging and not reinforcing an easy behavior, then asking for a difficult, time consuming behavior will increase the likelihood of a negative response because the animal must expending a lot more effort and wait a longer period of time before being reinforced.)
2) When you've reached the point where the animal can accept these simple non-reinforcement situations without reacting negatively, increase the number of non-reinforced behaviors. Then increase the difficulty of non-reinforced behaviors.
3) GO SLOW! Like desensitization, this is one of those types of training in which only the animal can determine the correct pace. Watch for negative reactions and, if you see anything you don't like, immediately regress. Keep in mind that this process is not a race. The idea is to come out with the result you want (acceptance of variable reinforcement) and nothing else.
4) Keep track of what you're doing. Make a detailed record of which behaviors are not reinforced, when, by whom and what the reaction of the animal was. Vary the non-reinforced behaviors. Don't become predictable.
