Operant conditioning: reinforcement, punishment and shaping

Transcript

RyanHere's a story. Every Monday, Jordan's manager picks apart her weekly report in front of the whole team. One Monday she wakes up with a knot in her stomach and calls in sick. The dread lifts almost instantly.

SoheeAnd a few months later, Jordan is missing a lot of Mondays.

RyanWelcome to EPPP Prep, the podcast. I'm Ryan.

SoheeI'm Sohee. Today is operant conditioning, learning through consequences. And we'll take on a term the OpenStax psychology textbook says many people confuse: negative reinforcement.

RyanLast time was classical conditioning, where something in the environment triggers a reflex. In operant conditioning, organisms learn to associate a behavior and its consequence. A pleasant consequence makes the behavior more likely to be repeated.

SoheeB. F. Skinner saw a gap. Classical conditioning is limited to existing behaviors that are reflexively elicited. It doesn't account for new behaviors, like riding a bike.

RyanSo he built on Edward Thorndike's law of effect. Behaviors followed by satisfying consequences are more likely to be repeated. Behaviors followed by unpleasant consequences are less likely to be repeated.

SoheeSkinner tested this mainly with rats and pigeons, in an operant conditioning chamber that came to be called the Skinner box. A rat presses a lever, or a pigeon pecks a disk, for food.

RyanNow the four terms. Here's the key that unlocks them. Positive and negative don't mean good and bad. Positive means you're adding something. Negative means you're taking something away.

SoheeAnd reinforcement means the behavior increases. Punishment means it decreases.

RyanSo you get a two by two. Positive reinforcement adds something to increase a behavior. Negative reinforcement removes something to increase it. Positive punishment adds something to decrease a behavior. Negative punishment removes something to decrease it.

SoheePositive reinforcement first. A desirable stimulus is added to increase a behavior. The textbook calls it the most effective way to teach a person or an animal a new behavior.

RyanPaychecks and high grades are rewards. The textbook even describes second graders in Dallas who were paid two dollars each time they read a book and passed a short quiz on it. Reading comprehension rose significantly.

SoheeNegative reinforcement. An undesirable stimulus is removed to increase a behavior. The classic is the seatbelt system that beeps until you buckle. The annoying sound stops when you buckle, so you're more likely to buckle up in the future.

RyanHorse training too. A rider applies pressure with the reins or the legs, then removes the pressure when the horse turns or speeds up.

SoheeAnd here's the textbook flagging the confusion directly. Many people confuse negative reinforcement with punishment, but they're very different. Reinforcement, even when it's negative, always increases a behavior. Punishment always decreases one.

RyanSo, Jordan. What's growing? Calling in sick on Mondays. What happened right after? An unpleasant feeling, the dread, went away. Something removed, behavior up. Negative reinforcement.

SoheeThen punishment. Positive punishment adds an undesirable stimulus to decrease a behavior, like scolding a student to stop texting in class. Negative punishment removes a pleasant stimulus, like taking away a child's favorite toy after misbehavior.

RyanPause and think. A teenager breaks curfew, and her parents take her phone for the weekend. Her late nights drop off. Which of the four is this?

SoheeSomething pleasant, the phone, was taken away, and the behavior, breaking curfew, went down. That's negative punishment.

RyanNow, why are psychologists cautious about punishment, especially physical punishment? The textbook lists drawbacks. First, punishment may teach fear. A child may come to fear the parent who delivered it.

SoheeAnd children punished by teachers may come to fear the teacher and try to avoid school. Second, punishment may make children more aggressive and prone to antisocial behavior and delinquency. Children who see a parent spank when angry may act out the same behavior when they're angry.

RyanSkinner suggested weighing punishment against its possible negative effects. And today's psychologists and parenting experts favor reinforcement over punishment. Catch the child doing something good, and reward it.

SoheeBut what if the behavior you want never shows up on its own? For reinforcement to work, the organism must first display the behavior. So Skinner often used shaping: rewarding successive approximations of a target behavior.

RyanThe steps go like this. Reinforce any response that resembles the desired behavior. Then reinforce the response that more closely resembles it, and stop reinforcing the earlier one. Keep reinforcing closer and closer approximations. Finally, reinforce only the desired behavior.

SoheeSkinner used shaping to teach pigeons to turn in circles, walk in figure eights, and even play ping pong.

RyanNow, what counts as a reinforcer? Primary reinforcers have innate reinforcing qualities. They aren't learned. Water, food, sleep, shelter, sex and touch, among others. Pleasure, too.

SoheeA secondary reinforcer has no inherent value. It only reinforces when it's linked with a primary reinforcer. Praise is one. Money is another, worth something only because it buys other things. Stickers on a behavior chart, too.

RyanAnd tokens. Tokens are secondary reinforcers that can be traded in for rewards and prizes. Whole behavior management systems, called token economies, are built around them, and they've been found very effective at changing behavior in schools, prisons and mental hospitals.

SoheeThe textbook gives a classroom example. Young children earned a quiet hands token when they didn't hit or pinch, and lost a token when they did. They traded tokens for minutes of playtime.

RyanNotice there are two consequences packed in there. Earning a token adds something good, so that's positive reinforcement of keeping hands to yourself. Losing a token takes away something valued, so by our definitions, that's negative punishment of hitting.

SoheeAnd for any of this behavior modification to work, the textbook lists three conditions. The reinforcement has to be connected with the behavior, it has to matter to the child, and it has to be done consistently.

RyanTime-out deserves a careful look too, because it's easy to get wrong. The textbook says time-out operates on the principle of negative punishment. The child is removed from the desirable activity at hand. Here's when it fails. Make sure the child is removed from a desirable activity to a less desirable place. If the activity is something the child finds undesirable, time-out backfires, because being removed is more enjoyable.

SoheeThe rule of thumb is one minute for each year of the child's age. And during time-out, ignore the child, because caregiver attention may reinforce the misbehavior. Then a hug or a kind word when it's over.

RyanPause and think. A boy throws his pencil whenever the math worksheet comes out, and his teacher sends him to time-out in the hallway each time. The pencil throwing increases. What went wrong?

SoheeHe was removed from an activity he finds unpleasant. So time-out took away something aversive, and the throwing grew. Functionally that's negative reinforcement, not punishment. Exactly the backfire the textbook warns about.

RyanHere's where this gets clinical. The textbook describes a learning account of obsessive-compulsive disorder, often shortened to O C D, citing Mowrer. In that account, symptoms are learned responses, acquired and sustained through a combination of classical and operant conditioning.

SoheeFirst, classical conditioning. A neutral stimulus becomes associated with something that provokes anxiety or distress, so it starts triggering anxiety and obsessive thoughts. Then a ritual, repeated, reduces the anxiety.

RyanAnd efforts to relieve anxiety that way are negative reinforcement. Compulsive acts may be sustained because they reduce the anxiety triggered by the conditioned stimulus.

SoheeWhich is exactly the pattern in this site's practice item on negative reinforcement. A client leaves crowded stores when her heart starts racing, and her anxiety drops. A ten-year-old screams until his father puts the worksheet away. A driver's car chimes until the seat belt is fastened.

RyanAnd the study note for that item gives a two-question method. Did the behavior grow or shrink? Was something added or taken away? A behavior that grows because it ends something unpleasant is negative reinforcement.

SoheeAnd if the behavior grows, rule out both kinds of punishment. Punishment, positive or negative, always makes a behavior go down.

RyanLet's recap. Operant conditioning links a behavior to its consequence, building on Thorndike's law of effect.

SoheePositive adds, negative removes. Reinforcement increases, punishment decreases.

RyanNegative reinforcement is not punishment. Something unpleasant is removed, and the behavior goes up. Escape, like Jordan's Mondays, or a compulsive ritual that relieves anxiety.

SoheeShaping reinforces successive approximations. Primary reinforcers are innate. Secondary reinforcers, like praise, money and tokens, get their power from links to primary ones. Token economies work in schools, prisons and mental hospitals.

RyanTime-out is negative punishment, and it backfires when the child is escaping something unpleasant. And punishment has costs, so reinforcement is favored.

SoheeIf you've ever second-guessed yourself on negative reinforcement, you're in very good company. The textbook itself says many people confuse it. Now you have two questions that sort it out.

RyanTry a few practice questions on this one while it's fresh. You've got the tools. See you next time on EPPP Prep.

Loading the read-along timings…