Shaping

Design Pattern

• Break down target behavior. When engaging in the desired behavior (e.g. talking in front of a big crowd) is too overwhelming, then break it down into smaller bits that start small and progressively build up into more complex behavior that gradually resembles the target behavior.
• Successively approximate. Start small (e.g. just standing on stake, then saying hello) to successively approximate and build up to the final desired behavior.
• Use to increase or decrease target behavior. Introduce rewards to increase a behavior and punishments to decrease a behavior. A reward can both be positive (adding a favorable event or outcome) and negative (removing of unfavorable event or outcome). Similarly can punishments be both positive (add unfavorable event our outcome) or negative (remove favorable event or outcome).

Clearly define a target behavior with your users. Then design a program that rewards any response that in some way resembles the target behavior and gradually moves your reinforcement only to behavior that is closer to the target behavior until the final target behavior is achieved, which then is the only behavior to be rewarded.