Ch 6 - Operant Conditioning

Operant Conditioning

learning paradigm where a behavior is motivated by some kind of consequence

• behavior operates on the enviro/thing coming out of it

• behavior is instrumental in eliciting some response

• more intuitive to us to acheive a goal

• ex → train animals to do tasks, reward/punish people to influence beh

Thorndike’s Law of Effect

believed people were uncritically accepting anecdotes of animal intelligence

• experiment → tasked felines w/ escaping from a puzzle box by standing on a plate

• performance improved over time → only due to randomly successful actions

Law of Effect

behaviors leading to a desirable outcome are stamped in/reinforced while those leading to annoyance are stamped out/inhibited

• animal X understand problem it’s in or its solution → accidentally does beh

• reinforced → more common in animal’s brain

• inhibited → annoyance/failure to progress

Selection by Consequences

Skinner → wanted a controlled means of producing reliable behaviors

• originally thought behaviors were reflexes, saw them more voluntary

• operant chamber/skinner box

• novel/naive animal & thing to interact w/ → X guide, wait for them to interact

• animals in chamber allowed to freely behave

• X stim provided until did desired response

Skinner & Operant Conditioning

disapproved of Thorndike’s ideas of considering responses as desirable or annoying

• focused on whether behavior potentiality inc or dec

Parts of an Operant Paradigm

1. operant → response (R)/behavior

2. consequence/stim that affects probability of behavior occurring

3. discrim stim (SD) that signals whether a consequence will occur/suitable at time

light on (SD):lever press → food (SR/reinforcing stim)

Operants

emitted by organism, voluntary & flexible

• ex → more than 1 way to cook an egg

• interested in a class of responses

Class of Responses

group of behaviors achieving some goal

• exactly how they get there is variable

Reinforcers

consequences that in behavior probability

• positive or negative

Positive Reinforcer

adds an appetitive stim

• more likely to do behavior

• ex → giving your dog a treat

Negative Reinforcer

removes an aversive stim

• inc beh probability

• ex → making a child no longer grounded, remove grounding = inc probability to do desired beh

Punishers

consequences that dec behavior probability

• positive or negative

Positive Punisher

adds an aversive stimulus that decreases the likelihood of a behavior.

• ex → shouting no at a misbehaving dog

Negative Punisher

removes an appetitive stim

• ex → taking away a child’s favored toy

Interpreting Motivation/Consequence

actual motivation/consequence can sometimes be hard to interpret

What can negative reinforcement result in?

2 types of behavior → escape behaviors, avoidance behaviors

• removing something reliable to make beh happen more

Escape Behaviors

done to remove the aversive stim, comes from negative reinforcement

Avoidance Behaviors

occur before any aversive stim & prevent it from being experienced at all

• results from negative reinforcement

How are positive & negative punishment used?

common when trying to affect behaviors of others

• more effective ways to do what we want

Interchanging Punishment & Reinforcement

effect of punishment can be a reinforcer for the punisher

• cat scratching furniture (R) → spraying cat w/ water (SP)

• quickly disciplining cats (R) → cat stops scratching (SR)

• could lead to a degrading relationship

• X making good feeling go away, just adding a punishment

Relationship as Example of Positive Punishment

she takes too long to pick magazine (R) → gets yelled at (S^p)

• him being aggressive/rude (R) → gets hug/affection (S^r)

• he orders her around (R) → finish selection (S^r)

• in a store (NS):male aggression (US) → submission/appeasement (UR)

  • in a store (CS) → appease (CR)/affection (R) → guy X mean (S^r)

What is the response & S^r?:

A Dog won’t stop barking. FInally to quiet them down, you give them a chew to keep them busy.

Dog barks (R) → get treat (S^r)

• actually reinforcing the barking

What is the response & S^r?:

If you don’t get a B+ average, you won’t be given a paid vacation.

get a B+ average (R) → get vacation (S^r)

• you can’t have a contingency for a non-action

What is the response & S^r?:

If you see the traffic light turn red, you will hit the brakes.

red light (S^D): step on brake (R) → avoid getting a traffic ticket (S^r)

• hit brakes to avoid punishment of not doing so

When do reinforcers work best?

if conseq is immediate, w/in </= 30s

• effectiveness drops framatically after this

Primary Reinforcers

inherently appetitive stimuli → food, comfort, sexual activity, etc

• stim may be a good reinforcer at 1 moment but ineffective in another moment

Secondary Reinforcers

learned to be appetitive by association w/ a primary reinforcer or other event

Classically Conditioning Reinforcers

tone (NS):food (US) → salivation (UR)

tone (CS) → salivation CR

lever press (R) → tone (S^r)

How can discriminating stimuli also be reinforcers?

Discriminative stimuli signal the availability of reinforcement, making them effective in influencing behavior. When associated with rewards, they become secondary reinforcers that encourage certain responses.

• light (S^d):lever press (R) → tone (S^r)

Generalized Reinforcers

secondary reinforcers that have been associated w/ many primary reinforcers (ex → money = comfort, food, sex)

Intrinsic Reinforcement

comes w/ internal goals or motivations

• doing something because it feels good is a typical indicator

Extrinsic Reinforcement

comes from an external source or due to an external motivation

• line b/w ext & int X always clear, value can change over time

  • ex → you are good at knitting & start selling creations, becomes work

Natural Reinforcers

typical conseqs of a behavior

• typically superior to contrived, produce more reliable behavior change

Contrived Reinforcers

introduced explicitly to cause a behavioral change

Shaping

The process of reinforcing successive approximations toward a desired behavior

• training complex behs requires multiple trials

• reward behavior that’s progressively closer to goal behavior