Operant Conditioning and Reinforcement Schedules

Recap of Previous Lecture

  • Learning definition: Relatively long-lasting changes in behavior caused by environmental events, excluding simple effects of maturation, aging, surgery, or drugs.
  • Justification for using animals:
    • Core reason: Phylogenetic continuity, or evolutionary reasons. Human and animal behavior are made up of similar simple elements, and the way the environment controls behavior is the same for both.
    • Assumption: Behavior is controlled by its consequences. Reinforced behaviors increase; non-reinforced or punished behaviors decrease.
  • Thorndike and the Law of Effect:
    • Experiment: Cat in a puzzle box.
    • Law of Effect: If a response is followed by a favorable consequence, it's likely to be repeated, strengthened, or reinforced. The favorable consequence is called a reinforcer.
  • Shaping:
    • Used when the desired behavior never occurs.
    • Method of successive approximations: Reinforce closer and closer approximations to the target behavior.

Shaping a pigeon – BF Skinner

  • BF Skinner shaped the pigeon to turn circles.
  • Skinner Box: Highly controlled environment. The rate at which a pigeon pecked could be precisely measured and controlled by changing the consequences. Skinner could present stimuli, record responses, and arrange schedules of reinforcement without a researcher being present.
  • Operant conditioning: The change that takes place when consequences have a particular effect on behavior.
  • Operant behavior: Behavior that operates on the environment and produces consequences.

Operant behavior

  • Controlled by its consequences.
  • Presentation of a favorable or appetitive event (something desirable).
  • Positive reinforcement: Presenting something; positive equals addition.
  • Negative reinforcement: Something being removed or withdrawn; negative means subtraction.
  • Reinforcement: Increases behavior.
  • Punishment: Decreases behavior.
Two by two matrix
  • Event presented: Positive contingency.
  • Event withdrawn: Negative contingency.
  • Appetitive event: Positive reinforcement (increases behavior).
  • Aversive event: Positive punishment (decreases behavior).
  • Negative reinforcement: Withdrawal of something aversive; removing the aversive stimulus, which increases behavior.
The Law of Effect Should Be Considered In Two Parts
  • If a response is followed by a favorable consequence (presentation of something nice or removal of something nasty), it will be more likely to happen again.
  • If a response is followed by an unfavorable consequence (presentation of something nasty or removal of something appetitive), it will be less likely to happen again.
Extinction
  • If a response used to be followed by reinforcement, and now the reinforcement stops, it's called extinction.
  • Behavior is less likely to happen again.
Analyzing Contingencies
  • Analyze situations according to the contingencies that are operating to predict behavior.
  • Sending someone to prison:
    • Society's intention: Punishment to reduce criminal behavior.
    • Negative punishment: Withdrawal of freedom.
    • Positive punishment: Adding unpleasant environment.
    • Possible positive reinforcement: Access to food and television.
    • Possible negative reinforcement: Removal of difficult responsibilities.
    • If prison is better than the outside world, it could increase offending.
The Dog Runs Away Example
  • Dog runs away, owner says, "Come back, Rover," and then smacks the dog.
  • Mistaken intention: Punishing running away.
  • Actual consequence: Punishing returning.
  • Result: Dog will be more likely to stay away longer.
Punishment Works Best
  • If it's immediate and introduced suddenly rather than faded in slowly.
  • Sudden vs. graded introduction of punishment (Ezrin, 1960s):
    • Rats pressing a lever for food, then punished with electric shocks.
    • Sudden introduction of moderate shock eliminated behavior immediately.
    • Gradual increase of shock intensity resulted in needing much more severe punishment to eliminate behavior.
  • Boiling Frog example:
    • Gradual heating of water won't cause a frog to jump out, but sudden immersion in boiling water will.
  • Reinforcement needs a relatively big change in the environment to produce a big change in behavior.
  • Much of behavior is controlled by avoiding punishment rather than receiving positive reinforcement.
Real Life example
  • Kids are fighting, and I turn on the TV to keep them quiet.
  • My behavior (turning on TV) is negatively reinforced (removal of aversive event).
  • Kids' fighting behavior is positively reinforced (presentation of TV).
  • Feedback cycle: Fighting leads to TV, creating a circular relationship.
  • Reinforcement works when pigeon does what he wanted them to do.
  • Control is in the environment; it's always reciprocal, with feedback loops.
  • Behavior does not change for behaviour's sake
    *
Primary and Secondary Reinforcers
  • Primary reinforcers: Innate biological significance, important to survival (e.g., food, warmth, water, escape from pain).
  • Secondary reinforcers: Previously neutral stimuli that have acquired reinforcing properties through pairing with primary reinforcers (also called conditioned reinforcers).
  • Examples:
    • Light paired with food for pigeons.
    • Saying "good dog" (paired with primary reinforcement in the past).
  • Most behavior is maintained by secondary reinforcers.