The action then leads to a pleasurable event, but only sometimes, and this is step 3: a variable reward. Maybe she’ll find some expression of praise or friendship, maybe not. This is a key discovery of behaviorist psychology: It’s best not to reward a behavior every time the animal does what you want. If you reward an animal on a variable-ratio schedule (such as one time out of every 10 times, on average, but sometimes fewer, sometimes more), you create the strongest and most persistent behavior.
