Behaviorists have identified a variety of stimuli and events that can reinforce and thereby increase learners’ behaviors. They distinguish between two general categories of reinforcers: primary ver- sus secondary. They also suggest that reinforcement can take either of two forms: positive or negative.
Primary versus Secondary Reinforcers
A primary reinforcer satisfies a built-in, perhaps biology-based, need or desire. Some primary reinforcers are essential for physiological well-being; examples are food, water, oxygen, and warmth. Others enhance social cohesiveness and so indirectly enhance one’s chances of survival; examples are physical affection and other people’s smiles (Harlow & Zimmerman, 1959; Vollmer & Hackenberg, 2001). To some degree, there are individual differences regarding the extent to which certain consequences serve as primary reinforcers; for instance, sex is reinforcing for some individuals but not others, and a particular drug may be a primary reinforcer for a drug addict but not necessarily for a nonaddicted individual (e.g., Lejuez, Schaal, & O’Donnell, 1998).
A secondary reinforcer , also known as a conditioned reinforcer , is a previously neutral stimu-
lus that has become reinforcing to a learner through repeated association with another reinforcer. Examples of secondary reinforcers, which don’t satisfy any built-in biological or social needs, are praise, good grades, and money. 2
How do secondary reinforcers take on reinforcing value? One early view involves classical conditioning: A previously neutral stimulus is paired with an existing reinforcer (UCS) that elic- its a feeling of satisfaction (UCR) and begins to elicit that same sense of satisfaction (CR) (Bersh,
2 One could argue that praise enhances social relationships and so might be a primary reinforcer. However, praise involves language—a learned behavior—and not all individuals find it reinforcing, hence its catego- rization as a secondary reinforcer (e.g., see Vollmer & Hackenberg, 2001).
1951; D’Amato, 1955). An alternative perspective is that secondary reinforcers provide informa- tion that a primary reinforcer might subsequently be coming (G. H. Bower, McLean, & Meachem, 1966; Fantino, Preston, & Dunn, 1993; Green & Rachlin, 1977; Mazur, 1993). The second explanation has a decidedly cognitive flavor to it: The learner is seeking information about the
environment rather than simply responding to that environment in a “thoughtless” manner. The relative influences of primary and secondary reinforcers in our lives probably depend a great deal on economic circumstances. When such biological necessities as food and warmth are scarce, these primary reinforcers, as well as secondary reinforcers closely associated with them (e.g., money), may be major factors in reinforcing behavior. But in times of economic well-being, when cupboards are full and houses are warm, such secondary reinforcers as praise and good grades are more likely to play major roles in the learning process.
Positive Reinforcement
When people think about stimuli and events that are reinforcing for them, they typically think about positive reinforcement. In particular, positive reinforcement involves the presentation of a
stimulus after the response. Positive reinforcement can take a variety of forms: Some are extrin-
sic reinforcers , in that they’re provided by the outside environment, whereas others come from within the learner.
Material reinforcers A material reinforcer (also known as a tangible reinforcer ) is an actual object; food and toys are examples. Material reinforcers can be highly effective in changing behavior, especially for animals and young children. However, most psychologists recommend that at school, teachers use material reinforcers only as a last resort, when absolutely no other reinforcer works. Food, toys, trinkets, and similar items have a tendency to distract students from the things they should be doing in class and thus may be counterproductive over the long run.
Social reinforcers A social reinforcer is a gesture or sign (e.g., a smile, attention, or praise) that one person gives another, usually to communicate positive regard. In classroom settings, teacher attention, approval, and praise can be powerful reinforcers (P. Burnett, 2001; McKerchar & Thompson, 2004; N. M. Rodriguez, Thompson, & Baynham, 2010). 3 Attention or approval
from peers can be highly reinforcing as well (F. E. Bowers, Woods, Carlyon, & Friman, 2000; Flood et al., 2002; Grauvogel-MacAleese & Wallace, 2010).
Activity reinforcers Speaking nonbehavioristically, an activity reinforcer is an opportunity to engage in a favorite activity. (Quick quiz: Which word is the nonbehaviorist part of my defini- tion, and why?) David Premack (1959, 1963) discovered that people will often perform one activity if doing so enables them to perform another. His Premack principle for activity reinforc- ers is as follows:
When a normally high-frequency response follows a normally low-frequency response, the high-frequency response will increase the frequency of the low-frequency response. A high-frequency response is, in essence, a response that a learner enjoys doing, whereas a low- frequency response is one that the learner doesn’t enjoy. Another way of stating the Premack
3 We should note here that praise has potential downsides, depending on the specific message it conveys and the context in which it’s given; we’ll examine its possible downsides in Chapters 16 and 17 .
principle, then, is that learners will perform less-preferred tasks so that they can subsequently engage in more-preferred tasks.
To illustrate, I rarely do housework; I much prefer retreating to my home office to either read or write about human learning and behavior. I’ve found that I’m more likely to do household chores if I make a higher-frequency behavior, such as reading a mystery novel or hosting a party, contingent on doing the housework. In a similar way, appropriate classroom behavior can be improved through the Premack principle. For instance, young children can quickly be taught to sit quietly and pay attention if they’re allowed to engage in high-frequency, high-activity behaviors (e.g., interacting with classmates) only after they’ve been quiet and attentive for a certain period of time (Azrin, Vinas, & Ehle, 2007; Homme, deBaca, Devine, Steinhorst, & Rickert, 1963).
Positive feedback In some instances, material and social reinforcers improve classroom behavior and lead to better learning of academic skills because they communicate a message that learners are performing well or making significant progress. Such positive feedback is clearly effective in bringing about desired behavior changes (Kladopoulos & McComas, 2001; S. Ryan, Ormond, Imwold, & Rotunda, 2002; Shute, 2008; R. E. Smith & Smoll, 1997).
I once spent a half hour each day for several weeks working with Michael, a 9-year-old boy with a learning disability who was having difficulty learning cursive letters. In our first few ses- sions together, neither Michael nor I could see any improvement, and we were both becoming increasingly frustrated. To give ourselves more concrete feedback, I constructed a chart on a sheet of graph paper and explained to Michael how we would track his progress by marking off the number of cursive letters he could remember each day. I also told him that as soon as he had reached a dotted line near the top of the page (a line indicating that he had written all 26 letters correctly) for three days in a row, he could have a special treat (his choice was a purple felt-tip pen). Michael’s daily performance began to improve dramatically. Not only was he making noticeable progress but he also looked forward to charting his performance and seeing a higher mark each day. Within two weeks, Michael had met the criterion for his felt-tip pen: He had writ- ten all 26 cursive letters for three days in succession. As it turns out, the pen probably wasn’t the key ingredient in our success: Michael didn’t seem bothered by the fact that he lost it within 24 hours of earning it. Instead, I suspect that the concrete positive feedback about his own improve- ment was the true reinforcer that helped Michael learn.
Feedback is especially likely to be effective when it communicates what students have and haven’t learned and when it gives them guidance about how they might improve their perfor- mance (Hattie & Timperley, 2007; Shute, 2008). Under such circumstances, even negative feed-
back can lead to enhanced performance. It’s difficult to interpret this fact within a strictly behaviorist framework; it appears that students must be thinking about the information they
receive and using it to modify their behavior and gain more favorable feedback later on.
Intrinsic reinforcers Oftentimes learners engage in certain behaviors not because of any external consequences but because of the internal good feelings—the intrinsic reinforcers —that such responses bring. Feeling successful after solving a difficult puzzle, feeling proud after returning a valuable item to its rightful owner, and feeling relieved after completing a difficult assignment are all examples of intrinsic reinforcers. People who continue to engage in responses for a long time without any obvious external reinforcers for their efforts are probably working for intrinsic sources of satisfaction.
The concept of intrinsic reinforcers doesn’t fit comfortably within traditional behaviorism, which, as you should recall, focuses on external, observable events. Yet, for many students, the
true reinforcers for learning are probably the internal ones—the feelings of success, mastery, and pride—that their accomplishments bring. For such students, other reinforcers are more helpful if they provide feedback that academic tasks have been performed well. Grades may be reinforc- ing for the same reason: Good grades reflect high achievement—a reason to feel proud.
Positive feedback and the intrinsic reinforcement that such feedback brings are probably the most productive forms of reinforcement in the classroom. Yet teachers must remember that what is reinforcing for one student may not be reinforcing for another; reinforcement, like beauty, is in the eyes of the beholder. In fact, consistent positive feedback and resulting feelings of success and mastery can occur only when instruction has been carefully tailored to individual skill levels and abilities and only when students have learned to value academic achievement. When, for what- ever reasons, students aren’t initially interested in achieving academic success, then other reinforc- ers, such as social and activity reinforcers—even material ones if necessary—can be useful in helping students acquire the knowledge and skills that will serve them well in the outside world.
Negative Reinforcement
In contrast to positive reinforcement, negative reinforcement increases a response through the
removal of a stimulus, usually an aversive or unpleasant one. Don’t let the word negative lead you
astray here. It’s not a value judgment, nor does it mean that an undesirable behavior is involved; it simply refers to the fact that something is being taken away from the situation. For example,
imagine a rat in a Skinner box that often gives the rat an unpleasant electric shock. When the rat discovers that pressing a bar terminates the shock, its bar-pressing behavior increases consider- ably. Similarly, some cars emit an annoying sound if the keys are still in the ignition when the driver’s door is opened; removal of the keys from the ignition is negatively reinforced (and so will presumably increase in frequency) because the sound stops.
Removal of guilt or anxiety can be an extremely powerful negative reinforcer for human beings. A child may confess to a crime committed days or weeks earlier because she feels guilty about the transgression and wants to get it off her chest. Anxiety may drive one student to complete a term paper early, thereby removing an item from his things-to-do list. Another student confronted with the same term paper might procrastinate until the last minute, thereby removing anxiety—if only temporarily—about the more difficult aspects of researching for and writing the paper.
Negative reinforcement probably explains many of the escape behaviors that humans and
nonhumans learn. For instance, rats that are given a series of electric shocks will quickly learn to turn a wheel that enables them to escape to a different, shock-free environment (N. E. Miller, 1948). Likewise, children and adolescents acquire various ways of escaping unpleasant tasks and situations in the classroom and elsewhere. Making excuses (“My dog ate my homework!”) and engaging in inappropriate classroom behaviors provide means of escaping tedious or frustrating academic assignments (Dolezal & Kurtz, 2010; A. W. Gardner, Wacker, & Boelter, 2009; McKerchar & Thompson, 2004; Romaniuk et al., 2002). Lying about one’s own behaviors (“I didn’t do it— he did!”) is potentially a way of escaping the playground supervisor’s evil eye.
Complaints about nonexistent stomachaches, chronic truancy, and dropping out of school are ways of escaping the school environment altogether. Some escape responses are productive ones, of course; for instance, many teenagers acquire tactful ways of rebuffing unwanted sexual advances or leaving a party where alcohol or drugs are in abundance.
Keep in mind that negative reinforcement can affect the behavior of teachers as well as the behavior of students. Teachers often behave in ways that get rid of aversive stimuli; for example,
they may use classroom discipline strategies (responses such as yelling at students or promising less homework) that eliminate unpleasant stimuli (disorderly conduct) over the short run but are ineffective over the long run. As an illustration, if Ms. Jones yells at Marvin for talking too much, Marvin may temporarily stop talking, which negatively reinforces Ms. Jones’s yelling behavior. 4
But if Marvin likes getting Ms. Jones’s attention (a positive reinforcer for him), he will be chatter- ing again before very long.
Let’s return to the title of this chapter: “Instrumental Conditioning.” Instrumental condition- ing encompasses not only the behavior-increasing effect of reinforcement—an effect that Skinner called operant conditioning —but also the behavior-suppressing effect of punishment. We turn to
punishment now.
I can't — I have to wash my hair tonight.
I didn't do it — he did! I don't feel so good. I think I hear my mother calling.
4 Although behaving in a disruptive fashion is a response that Marvin makes, in this situation it serves as a stimulus that Ms. Jones wants to eliminate.
Common escape responses
Punishment
Most behaviorists define punishment in terms of its effect on behavior: It decreases the fre- quency of the response it follows. Punishment can take either of two forms. Punishment I involves the presentation of a stimulus, typically an aversive one—perhaps a scolding or a failing
grade. Punishment II involves the removal of a stimulus, usually a pleasant one; examples are
monetary fines for misbehaviors (because money is being taken away) and loss of privileges. Figure 4.3 illustrates the differences among positive reinforcement, negative reinforcement, Punishment I, and Punishment II.
Presented after the response Removed after the response Pleasant Aversive Positive Reinforcement (response increases) Punishment II (response decreases) Punishment I (response decreases) Negative Reinforcement (response increases) Nature of Stimulus Stimulus is Figure 4.3
Many people mistakenly use the term negative reinforcement when they’re really talking about
punishment. Although both phenomena may involve aversive stimuli, they differ in two critical respects. First, as you’ve seen, they have opposite effects: Negative reinforcement increases the
frequency of a response, whereas punishment decreases it. A second crucial difference concerns
the order of events. With negative reinforcement, the aversive stimulus stops when the response
is emitted. With Punishment I, however, the aversive stimulus begins when the response is emitted.
Figure 4.4 illustrates this difference graphically. The termination of an aversive stimulus negatively reinforces a response; the initiation of an aversive stimulus punishes a response.
Early research by both Thorndike (1932a, 1932b) and Skinner (1938) indicated that pun- ishment was unlikely to reduce the behavior it followed. For instance, Skinner (1938) found that when rats were punished for a response that had previously been reinforced, the response was temporarily suppressed but soon returned to its prepunishment frequency. But later research revealed that punishment can be effective in many situations. As a result, many behaviorists have
revived the “punishment” part of Thorndike’s original law of effect, asserting that responses fol- lowed by an unpleasant state of affairs are, in fact, weakened (e.g., Conyers et al., 2004; Lerman & Vorndran, 2002; G. C. Walters & Grusec, 1977).
Oftentimes punishment decreases behaviors very quickly. For example, in one study (R. V. Hall et al., 1971, Experiment 1), punishment virtually eliminated the aggressive behavior of 7-year-old Andrea, who was deaf. Initially, Andrea often pinched and bit both herself and any- body else with whom she came in contact; the frequency of such responses (an average of 72 per school day) was so high that normal academic instruction was impossible. Following a period of data collection without any intervention (a baseline period), punishment for each aggressive act
began: Whenever Andrea pinched or bit, her teacher pointed at her sternly and shouted “No!” Figure 4.5 shows the changes in Andrea’s behavior. (The brief reversal to a nonreinforcement
baseline period on day 25 was used to minimize the likelihood that other factors were responsi- ble for the behavior change.) Even though Andrea was deaf, the shouting and pointing virtually eliminated her aggressiveness.