Voluntary Delay of Reinforcement as a Function of Model Status.