6.2: Basics of Operant Conditioning

What is Operant Conditioning?

Before jumping into a lot of terminology, it is important to understand what operant conditioning is or attempts to do. But before we get there, let’s take a step back. So what happens when we make a behavior? Consider this framework:

Stimulus, also called an antecedent, is whatever comes before the behavior, usually from the environment. Response is a behavior. And of course, consequence is the result of the behavior that makes a behavior more or less likely to occur in the future. Presenting this framework is important because operant conditioning as a learning model focuses on the person making some response for which there is a consequence. As we learned from Thorndike’s work, if the consequence is favorable or satisfying, we will be more likely to make the response again (when the stimulus occurs). If not favorable or unsatisfying, we will be less likely. Recall that respondent or classical conditioning, which developed thanks to Pavlov’s efforts, focuses on stimulus and response, and in particular, the linking of two types of stimuli – NS and US.

Before moving on let’s state a formal definition for operant conditioning:

Return to our discussion of operant behaviors from Section 6.1.2. Operant behaviors are voluntary and controlled by their consequences. Let’s say a child chooses to study and does well on an exam. His parents celebrate the grade by taking him out for ice cream (the positive outcome of the behavior of studying). In the future, he studies harder to get the ice cream. But what if his brother decides to talk back to his parents and is scolded for doing so. He might lose television privileges. In the future, the brother will be less likely to talk back and stay quiet to avoid punishment. In both cases, the behavior was freely chosen and the likelihood of making that behavior in the future was linked to the consequence in the present. These operant behaviors are sometimes referred to as operants.

Operant behaviors are emitted by the organism such as in the case of the child studying or the child talking back. In contrast, respondent behaviors are elicited by stimuli (either the US or CS) such as a dog salivating to either the sight of food (US) or the sound of a bell (CS/NS). Also, operant behaviors can be thought of as a class of responses, all of which are capable of producing a consequence. For instance, studying is a general behavior, but it could include spaced study sessions such as spending 30 minutes after each class going over what was covered that day or massed studying the night before the exam. Both are studying behavior and it is easier to predict their occurrence than an exact type of behavior. In the case of talking back, consider that a child could just make a snide comment at the parent or engage in argumentative behavior. The result is the same (negative consequence) for engaging in this behavior, whatever form it may take exactly (the topography of the behavior).

Behavioral Contingencies

As we have seen, the basis of operant conditioning is that you make a response for which there is a consequence. Based on the consequence you are more or less likely to make the response again. Recall that a contingency is when one thing occurs due to another. Think of it as an If-Then statement. If I do X, then Y will happen. For operant conditioning, this means that if I make a behavior, then a specific consequence will follow. The events (response and consequence) are linked in time (more on this in a bit).

What form do these consequences take? There are two main ways they can present themselves.

• Reinforcement — Due to the consequences, a behavior/response is more likely to occur in the future. It is strengthened.
• Punishment — Due to the consequence, a behavior/response is less likely to occur in the future. It is weakened.

Reinforcement and punishment can occur as two types — positive and negative. These words have no affective connotation to them meaning they do not imply good or bad, or an emotional state. Positive means that you are giving something — good or bad. Negative means that something is being taken away — good or bad. Check out Table 6.1 below for how these contingencies are arranged. Bear in mind that the actual contingency presented in the cells needs to be inferred from what is along the outside. We know giving is positive and if you are giving some bad or aversive thing, you should be trying to weaken a behavior, which is punishment. Hence, giving something bad that would weaken a behavior is positive punishment. The others work the same. This is one way to teach this concept. If you learned another way and understand the four contingencies, stick with what you learned.

Table 6.1. Contingencies in Operant Conditioning

Let’s go through each:

• Positive Punishment (PP) — If something bad or aversive is given or added, then the behavior is less likely to occur in the future. If you talk back to your mother and she slaps your mouth, this is a PP. Your response of talking back led to the consequence of the aversive slap being delivered or given to your face.
• Positive Reinforcement (PR) — If something good is given or added, then the behavior is more likely to occur in the future. If you study hard and earn, or are given, an A on your exam, you will be more likely to study hard in the future. Your parents may also give you money for your efforts. Hence the result of studying could yield two PRs — the ‘A’ and the money.
• Negative Reinforcement (NR) — This is a tough one for students to comprehend because the terms do not seem to go together and are counterintuitive. But it is really simple and you experience NR all the time. This is when something bad or aversive is taken away or subtracted due to your actions, making you more likely to do the same behavior in the future when some stimulus presents itself. For instance, what do you do if you have a headache? You likely answered take Tylenol. If you do this and the headache goes away, you will take Tylenol in the future when you have a headache. NR can either result in current escape behavior or future avoidance behavior. What does this mean? Escape behavior occurs when we are presently experiencing an aversive event and want it to end. We make a behavior and if the aversive event, such as withdrawal symptoms from not drinking coffee for a while, goes away, then we can say we escaped the aversive state. Feeling hungry and eating food is another example. Taking Tylenol helps us to escape a headache. In fact, to prevent the symptoms from ever occurring, we might drink coffee at regular intervals throughout the day. This is called avoidance behavior. By doing so we have removed the possibility of the aversive event occurring and this behavior demonstrates that learning has occurred. We might also eat small meals throughout the day to avoid the aversive state of being hungry.
• Negative Punishment (NP) — This is when something good is taken away or subtracted making a behavior less likely in the future. If you are late to class and your professor deducts 5 points from your final grade (the points are something good and the loss is negative), you will hopefully be on time in all subsequent classes.

To easily identify contingencies, use the following three steps:

1. Identify if the contingency is positive or negative. If positive, you should see words indicating something was given, earned, or received. If negative, you should see words indicating something was taken away or removed.
2. Identify if a behavior is being reinforced or punished. If reinforced, you will see a clear indication that the behavior increases in the future. If punished, there will be an indication that the behavior decreases in the future.
3. The last step is easy. Just put it all together. Indicate first if it is P or N, and then indicate if there is R or P. So you will have either PR, PP, NR, or NP. Check above for what these acronyms mean if you are confused.

Primary vs. Secondary (Conditioned)

The type of reinforcer or punisher we use is important. Some are naturally occurring while some need to be learned. We describe these as primary and secondary reinforcers and punishers. Primary refers to reinforcers and punishers that have their effect without having to be learned or are unconditioned. Food, water, temperature, and sex, for instance, are primary reinforcers while extreme cold or hot or a punch on the arm are inherently punishing. A story will illustrate the latter. When I was about 8 years old, I would walk up the street in my neighborhood saying, “I’m Chicken Little and you can’t hurt me.” Most ignored me but some gave me the attention I was seeking, a positive reinforcer. So, I kept doing it and doing it until one day, another kid was tired of hearing about my other identity and punched me in the face. The pain was enough that I never walked up and down the street echoing my identity crisis for all to hear. Pain was a positive punisher and did not have to be learned. That was definitely not one of my finer moments in life.

Secondary or conditioned reinforcers and punishers are not inherently reinforcing or punishing but must be learned. An example was the attention I received for saying I was Chicken Little. Over time I learned that attention was good. Other examples of secondary reinforcers include praise, a smile, getting money for working or earning good grades, stickers on a board, points, getting to go out dancing, and getting out of an exam if you are doing well in a class. Examples of secondary punishers include a ticket for speeding, losing television or video game privileges, being ridiculed, or a fee for paying your rent or credit card bill late. Really, the sky is the limit, especially with reinforcers.

A particular type of secondary reinforcer is called a generalized reinforcer (Skinner, 1953) and obtains the name because of being paired with many other reinforcers. Consider the example of money. With it, we can purchase almost anything. Maybe we should even add credit or debit cards to this discussion since most of us carry them and not actual money nowadays. In Module 7 we will discuss applications of operant conditioning, one of which is behavior modification. In the token economy, generalized reinforcers are used in the form of tokens and are used to purchase items called backup reinforcers. More on this later.

Discriminative Stimuli

Sometimes a behavior is reinforced in the presence of a specific stimulus and not reinforced when the stimulus or antecedent is not present. These stimuli signal when reinforcement will occur, or not, and are called discriminative stimuli (also called an SD). Recall our earlier discussion of the S à R à C model or Stimulus-Response-Consequence. Though our focus was on R and C, discriminative stimuli show that S can be important too. Consider the rat in the Skinner box. Lever pushes earn food reinforcers, but what if this is true only if a light above the lever is on? In this case, the light serves as a discriminative stimulus and signals that reinforcement is possible. If the light is not on, reinforcement does not occur no matter how many times the lever is pushed (or how hard).

As a stimulus can signal reinforcement, so too it can signify punishment. At times my dog sprays the trashcan in an effort to mark his territory. When he does this he receives either a soft smack across the rear end or is verbally scolded. But if I see him approaching the trashcan, I can hold up my hand which usually signifies the punishment he will endure if he sprays the trashcan. The sight of my hand (or a newspaper rolled up) signals punishment if a behavior is engaged in.