psychology
Williams1990
chapter_6.pdf
6
Reinforcement Applications
Learning Objectives
After reading this chapter, you should be able to do the following:
• Describe how reinforcement principles can be applied to a variety of problematic behaviors, particularly in the areas of classroom management and autism.
• Explain the problem of extinction and discuss the factors that will maximize the chance of changes in behavior being maintained.
• Identify the potentially harmful effects of reinforcement, including the undermining of intrinsic motivation.
• Define and illustrate the “principle of minimal force.”
• Consider how the concept of self-control provides a possible alternative to willpower, how self- control principles can explain apparent failures of willpower, and how self-control can be taught.
• Understand how self-control principles can be used to increase the effectiveness of studying.
lie6674X_06_c06_201-230.indd 201 4/9/12 8:22 AM
CHAPTER 6Section 6.1 Three Applications
In the introduction to Chapter 5, we considered the paradox that the principles of rein- forcement appear so simple, yet are often remarkably difficult to apply in real life. One explanation, we suggested, was that the principles of reinforcement might not be as sim- ple as they first appear; in the course of that chapter, we encountered evidence supporting that view. Even brief delays of reinforcement, for example, have far more severe effects on learning than is commonly realized, and our understanding of phenomena such as motivation and stimulus control is still limited. Therefore, we could readily account for our difficulty in using reinforcement effectively by our incomplete understanding of the principles governing its use.
There is, however, another possibility. Even in those cases in which we do understand the principles of reinforcement, it can appear ineffective because we fail to apply the prin- ciples in a coherent and systematic way. We might already know enough to use reinforce- ment more effectively, if only we would apply that knowledge systematically. This, at any rate, was the belief of several learning psychologists in the 1960s, and it led to a major effort, under the rubric of behavior modification, to apply the principles of reinforce- ment developed in the animal laboratory to practical problems of human behavior. In this chapter we will review some of these programs, considering the extent to which they have been successful, and, insofar as they have failed, what these failures can tell us about the remaining gaps in our knowledge. We will begin by looking at some of the attempts that have been made to apply the principles of reinforcement to education.
6.1 Three Applications Reinforcement principles have been applied to a wide range of educational problems and institutional settings. We will focus our attention in this section on examples involving schoolchildren, teenage delinquents, and autistic children.
Classroom Behavior One of the most difficult problems for any teacher is children who are severely disruptive in class. By talking, moving around, and so on, they not only fail in their own learning but also seriously interfere with the work of those around them. To test a reinforcement-based program for dealing with this problem, Hall, Lund, and Jackson (1968) went to a school in a severely deprived urban area. They asked the teachers and principal to identify the chil- dren in the school whose behavior posed the most serious problems. One child, a third- grade boy named Robbie, had been in trouble ever since he entered the school. He had received repeated scoldings, been sent to the principal, and even been spanked—all to no avail. A classroom observer found that Robbie spent only 25% of his time on assigned tasks, with the remainder of his time devoted to activities such as talking, snapping rub- ber bands, drinking milk very slowly and then playing with the carton, and so on. His teacher often urged him to work; indeed, 55% of her contacts with Robbie occurred at times when he was not working.
If you were the teacher in charge of the class, what would you do? One natural reaction would be to punish him, but this had already been tried repeatedly without success. The experimenters’ analysis was that the teacher was actually encouraging Robbie’s misbehavior
lie6674X_06_c06_201-230.indd 202 3/14/12 4:38 PM
CHAPTER 6Section 6.1 Three Applications
by giving him attention when he misbehaved. As we saw in our discussion of social rein- forcement, attention from oth- ers, even when that attention comes in the form of scolding, can be reinforcing. The experi- menters therefore recommended that the teacher use attention to reinforce appropriate behavior. They asked her to ignore Robbie whenever he misbehaved. When he behaved appropriately for one minute, however, she was to come over and praise him, mak- ing comments such as, “Very nice, Robbie, you’ve been work- ing very well.”
The results of this intervention are shown in Figure 6.1. When reinforcement was introduced, there was an immediate increase in the proportion of time Robbie spent studying. When the teacher returned to the baseline condition—that is, scolding inappropriate behavior—studying fell; it improved again, however, when rein- forcement was reinstated. The purpose of alternating baseline and treatment phases is to assess whether any improvement is really the result of the treatment, or simply the pas- sage of time; this is called an ABA design. Moreover, this improvement proved durable: Observations made 14 weeks after the training program had ended revealed that Robbie was still spending 79% of his time working, compared with only 25% during the baseline phase. Not surprisingly, this change in the amount of time spent working also led to a substantial improvement in the quality of Robbie’s work. On spelling tests, for example, his performance improved from 57% to 97%.
Classroom mangement has always been a challenge for teachers, as demonstrated by the rowdy scene in this 19th century painting. Reinforcement that is both immediate and consistent has been shown to produce positive behavioral changes with disruptive students.
lie6674X_06_c06_201-230.indd 203 3/14/12 4:38 PM
CHAPTER 6Section 6.1 Three Applications
Figure 6.1: Effects of praise
Effect of praise on the proportion of class time Robbie devoted to studying. When he was praised as a reinforcer, Robbie spent significantly more time studying. Source: From Hall, R. V., Lund, D., & Jackson, D. (1968). Effects of teacher attention on study behavior. Journal of Applied Behavior Analysis, 1, 1–12, Figure 2. Reprinted by permission of John Wiley & Sons, Inc.
C la
ss T
im e
S p
e n
t S
tu d
y in
g (
p e
rc e
n t)
Class Sessions
5 10 15 20 25 30 35
0
100
80
60
40
20
baseline
reinforcement 1
reversal
reinforcement 2
postchecks
cueing discontinued
This improvement was not achieved without effort. To ensure that Robbie would be rein- forced immediately when he studied, his behavior had to be monitored constantly. To help the teacher, the classroom observer signaled surreptitiously whenever the criterion for reinforcement was met. In the early stages of the program, therefore, considerable effort was needed to implement it, but in the long term, the improvement in Robbie’s behavior meant that he required substantially less of the teacher’s attention, and this improvement was maintained when the observer’s signalling was discontinued. The experimenters obtained similar results with the other children studied. Thus, even a seemingly trivial reinforcer—just a little bit of praise and attention—produced remarkable changes in the behavior of the most severely disruptive children in the school, provided that this rein- forcement was both immediate and consistent.
lie6674X_06_c06_201-230.indd 204 6/5/13 11:38 AM
CHAPTER 6Section 6.1 Three Applications
The Token Economy This study supports the claim that many failures of reinforcement might be the result of reinforcement being used in a perfunctory, haphazard way. When reinforcement was used more systematically—and, in particular, when it was immediate and consistent—even a seemingly minor reinforcer, such as just a pat on the back, produced dramatic changes in behavior.
It would be misleading, however, to imply that applications of reinforcement principles are always this effective. The “reinforce good behavior, ignore bad behavior” strategy used by Hall, Lund, and Jackson has been found to be effective in many studies, but occasionally students whose disruptive behavior has been ignored have really gone hay-
wire, eventually forcing prema- ture cancellation of the program (for example, O’Leary, Becker, Evans, & Saudargas, 1969; see also the discussion of Patter- son et al.’s work in Chapter 7). In some of these failures, the problem was that the teachers did not implement the system properly. In the Hall, Lund, and Jackson study, for example, the technique was successful in six of the seven classes in which it was tried. In the one case where it failed, the teacher was unable to ignore the student’s bad behavior; the teacher continued to become angry and to scold the student whenever he was disobedient.
In other cases, the problem could have been that the social reinforcers used were not effec- tive reinforcers for the students concerned. As we noted in Chapter 5, social reinforcers gain or lose their effectiveness partly through experience; for some children, social rein- forcers such as praise and attention are not effective (at any rate, not when given by their teachers). Where praise fails, a possible alternative is the token economy, in which points or tokens are established as secondary reinforcers through pairings with a variety of more potent reinforcers. If children behave appropriately in class, for example, they are imme- diately given points that can later be exchanged for backup reinforcers such as candy. The advantages of using tokens as reinforcers are the following:
1. Because they are easily dispensed, they can be delivered immediately after the child makes a response.
2. Because they are exchangeable for a wide variety of backup reinforcers, they are always likely to be attractive. Even if a child does not want candy at a particular moment, the token might still be desirable because it can also be exchanged for other reinforcers such as toys.
With the right type of supportive services, individuals with developmental disabilities can be successfully employed, as shown by this image of disabled employees at a catering company. The developmentally disabled also respond well to token economy programs.
lie6674X_06_c06_201-230.indd 205 3/14/12 4:38 PM
CHAPTER 6Section 6.1 Three Applications
We have already encountered an example of a token economy in the study by Phillips (1968), in which juvenile delinquents were treated in a residential center called Achieve- ment Place. The boys were given points for appropriate behavior, and these points could then be exchanged for reinforcers such as snacks, money, permission to go into town, and so forth. When completion of homework assignments was reinforced with permission to stay up late, the average percentage completed was 50%; when points were used as the reinforcer for completion, this percentage rose to 100%.
Token economies have produced similar improvements in a wide range of settings. In one striking example, a token economy was used to reduce injuries in two mines (Fox, Hopkins, & Anger, 1987). Workers were given trading stamps at the end of each month in which they suffered no injuries, and they received extra stamps if their entire work group was injury-free. The stamps could be exchanged for a wide variety of items at redemption stores. Over a 12-year period, the number of injuries fell by 68% in one of the mines and by 85% in the other. The program was highly cost-effective for the owners: The cost of injuries fell by more than $260,000 a year, whereas the stamps cost only around $12,000. And the program also proved highly attractive to the workers: A union representative at one of the mines even asked that the token program be written into the workers’ contracts!
As we shall see, token economies need to be used with some caution. Nevertheless, they provide a potentially useful alternative for situations in which reinforcers such as praise prove ineffective.
Autism One of the most remarkable attempts to practically apply reinforcement principles has been the work of O. Ivar Lovaas with autistic children. Autism is a psychiatric disor- der in which children become totally isolated from their social environment, having very little verbal and physical contact with other human beings. At the severe end of the spec- trum, autistic children can spend their days rocking back and forth, sometimes engaging in bizarre and highly stereotyped gestures. Attempts to treat autism have largely been unsuccessful; in one long-term study of young children with this condition, more than 60% remained severely handicapped and had to be confined to hospi- tals (Rutter, 1970).
To treat this severely debilitat- ing condition, Lovaas devel- oped a program based almost entirely on the principles of reinforcement we have been dis- cussing. He viewed autism as a set of maladaptive behaviors and therefore set out to encour- age more appropriate behaviors using reinforcement, shaping, discrimination learning, and so
This image shows an autistic child in a psychologist’s office with his parents. Reinforcement programs based on modelling have been used effectively with autistic children.
lie6674X_06_c06_201-230.indd 206 3/14/12 4:38 PM
CHAPTER 6Section 6.1 Three Applications
on. To train autistic children to talk, for example, Lovaas used a shaping procedure similar to that described in Chapter 5 to train Dicky to wear his glasses. Children were immedi- ately reinforced with food whenever they made the desired response, and training started with simple responses such as pronouncing a single word. Their program was based on a technique called modelling, in which the teachers first demonstrated or modeled the desired behavior—in this case, pronouncing a word—and then reinforced the children as soon as they repeated it. Once the children had learned to pronounce the desired words, they were given discrimination training to help them learn how to use the words prop- erly—for example, they would be shown a toy and asked its name, with reinforcement given only if they gave the correct name.
Training was a demanding process, spread over many months, but the children did grad- ually make impressive progress (Lovaas, Koegel, Simmons, & Long, 1973). This prelimi- nary success allowed Lovaas to obtain funds to extend his program. His initial program had been restricted to a small group of children in a psychiatric hospital, but Lovaas was now able to train therapists to work with children in their own homes. The children were all under 4 years of age, and they were assigned therapists who worked with them 40 hours per week for approximately 2 years. The children’s parents were also given train- ing in the appropriate use of learning principles, so that appropriate behavior could be reinforced whenever it occurred.
To assess the effectiveness of the treatment program, Lovaas (1987) compared the behav- ior of the 19 children who participated with that of a control group who were either not treated or else treated only 10 hours per week. The results were quite remarkable. The IQ of the treated group increased by an average of 30 points compared with that of the controls, and 47% of the treated children improved sufficiently to be enrolled in public schools, where their behavior was indistinguishable from that of normal children. In con- trast, only 2% of children in the control conditions showed this level of improvement. Lovaas has prepared a film showing the behavior of the children before treatment and after, and the transformation in their behavior is so dramatic that it is sometimes hard to believe that you are seeing the same children. (Note that a DVD version of the film can be purchased at www.lovaas.com.)
The results reported by Lovaas and his colleagues were very, very impressive, but his methods proved controversial. One problem was cost: Treatment requires intensive tuition for an extended period and is thus very expensive. On the other hand, as Lovaas (1987) has pointed out, the cost of one full-time teacher for two years is approximately $40,000, in contrast to “the nearly $2 million incurred (in direct costs alone) by each client requir- ing life-long institutionalization.” Another problem concerned the use of punishment. In order to suppress behaviors that interfered with teaching, such as repetitive hand wav- ing, Lovaas made extensive use of punishment, typically in the form of shouting at the children when they misbehaved. This use of punishment proved intensely controversial, and it was eventually found that the treatment could be just as effective without it (e.g., Eikeseth, Smith, Jahr, & Eldevik, 2002).
A further, fundamental issue was whether Lovaas’ results could be replicated: Were they reliable, or just a fluke? Replication was not easy—the treatment is intensive and costly, and requires many years—but a number of replication studies have now been reported, and to an impressive extent the results have mirrored the outcomes achieved by Lovaas
lie6674X_06_c06_201-230.indd 207 3/14/12 4:38 PM
CHAPTER 6Section 6.2 The Problem of Maintaining Behavior
(for example, Birnbrauer & Leach, 1993; Sallows & Graupner, 2005). In an authoritative review published in the journal of the American Academy of Pediatrics, the Lovaas or ABA treatment was found to produce “substantial, sustained gains in IQ, language, aca- demic performance and. . .social behavior, and their outcomes have been significantly bet- ter than those of children in control groups” (Myers & Johnson, 2007; see also Virués- Ortega, 2010). The best outcomes have been obtained when treatment is intensive (40 hours per week) and starts at an early age (e.g., Howard, Sparkman, Cohen, Green, & Stanislaw, 2005 ). Under these conditions, it does look as if the Lovaas treatment has the potential to transform the lives of many children suffering from autism.
6.2 The Problem of Maintaining Behavior When psychologists first attempted to apply the principles of reinforcement to problem behaviors, there was considerable doubt that they would succeed. Could the behavior of delinquents, much less of children with severe learning disabilities or of psychotics, really be altered just by reinforcing appropriate behavior? Over the years, it has become clear that the answer is yes: Provided that reinforcement is used in a coherent and system- atic way, it can be effective in settings as diverse as elementary schools and universities, prisons and psychiatric wards. It is now well established that behavior can be altered by reinforcement; the greater problem has proved to be maintaining these gains when the reinforcement program is terminated.
The Problem of Extinction Consider, for example, the token economy used in Achievement Place. The initial results obtained with this program were highly positive. Boys who participated in this program were found to have substantially lower rates of court appearances during the two-year period following their participation than did boys with similar backgrounds, and par-
ticipants also had higher grades in school (Fixsen et al., 1978). When their behavior was exam- ined over a longer period, how- ever, much of this improvement was lost (Wolf, Braukmann, & Ramp, 1987). When you think about it, this is perhaps not sur- prising: If delinquents return to the environment that produced their delinquent behavior, it is understandable that they might return to the patterns of behavior they had previously found to be effective in those environments.
We can see evidence of this effect on a smaller scale in the
These young offenders sit in a juvenile prison in the Ukraine. Token economy programs have proven effective in modifying the behavior of delinquents while they are held in institutions; the greater problem has proved to be maintaining such improvements after they are released.
lie6674X_06_c06_201-230.indd 208 3/14/12 4:38 PM
CHAPTER 6Section 6.2 The Problem of Maintaining Behavior
program used in Achievement Place to encourage boys to study. As noted earlier, when points were made contingent on Tom’s completion of homework, the percentage of assignments completed rose to 100%. When points were eventually discontinued, how- ever, the percentage of completions fell back to zero. In other words, when Tom received a reward he valued for studying, he studied; when this reward was discontinued, he stopped studying. After all, why should anyone persist in a behavior if it no longer pro- duces reinforcement?
The implicit assumption in the programs we have been reviewing is that there are sources of reinforcement in the natural environment that will maintain the desired behavior if it can be established initially. A delinquent might need external incentives to learn to read, for example, but once the behavior is established, the inherent pleasure available from reading books, newspapers, and so on should maintain the behavior. In some cases, though, it can take time for these natural reinforcers to develop; if the reinforcement pro- gram is to be effective, it might be necessary to ensure that the reinforced behavior will continue long enough for the natural reinforcers to assume control.
Tactics for Encouraging Maintenance Several of the principles reviewed in the previous chapter can be used to encourage persis- tence. One such technique is partial reinforcement. As we noted in Chapter 5, the greater the intermittency of reinforcement during training, the longer behavior will persist after reinforcement is terminated. In most reinforcement programs, therefore, continuous rein- forcement is used to establish a behavior initially, but the frequency of reinforcement is progressively reduced as training continues.
Reinforcement in a Variety of Settings
Another technique for maximizing the persistence of behavior is to reinforce it in a variety of settings. According to the principle of stimulus control, reinforcement strengthens behav- ior most in the particular setting in which training is given. By reinforcing behavior in a variety of settings, we can increase the likelihood that it will generalize widely, and thus not extinguish immediately when conditions change (for example, on leaving the classroom in which training was given).
Fading
A third potentially useful tech- nique involves fading out the reinforcement program gradu- ally, rather than terminating it abruptly. In our discussion of shaping, we saw that a reinforced
This two-year-old black Labrador is being trained at a special canine academy using positive reinforcement techniques. The behavioral changes will generalize if the training is provided not just at the academy but in a variety of different settings, such as at the dog park, inside the home, and so on.
lie6674X_06_c06_201-230.indd 209 3/14/12 4:38 PM
CHAPTER 6Section 6.2 The Problem of Maintaining Behavior
response is more likely to persist if any changes in the reinforce- ment program—for example, in the number of responses required for reinforcement— are introduced gradually rather than abruptly, and this principle also applies to the termination of the program. In a study by Hall and colleagues (1972), for exam- ple, the experimenters were stu- dents in a university course on behavior modification. These students carried out projects in their own homes using the principles studied in the course. One such project involved a boy named Jerry, who had started wearing an orthodontic device
when he was eight years old. Jerry was supposed to wear the device for 12 hours a day. In practice, though, he wore it for only a few hours a day because he hated it. After eight years, four dentists, and $3,300 in bills, Jerry’s condition was essentially unchanged.
As a first step toward altering this behavior, Jerry’s mother began to keep careful records of how often he wore the device, so she could accurately assess the effects of any treat- ment. During this baseline period, Jerry wore the device only 25% of the time (see Figure 6.2). To increase this percentage, his mother first tried social reinforcement. She did not reprimand her son when he failed to wear the device, but she praised him when he did. This social reinforcement produced a substantial increase in the desired behavior—he wore the device 36% of the time—but for practical purposes the increase was not suf- ficient. In the next phase, therefore, his mother tried a more powerful reinforcer: money. If Jerry was wearing the device when his mother checked, he received 25 cents; if he was not, he lost 25 cents. His mother paid him at the end of each month, and the amount of time Jerry spent wearing the device increased to 60%. To increase it still further, his mother changed to immediate reinforcement—Jerry received payment immediately after each inspection—and the wearing time now rose to 95%! As we have seen again and again (and again . . . ), a reinforcer presented immediately is generally far more effective than the same reinforcer presented after a delay.
Positive reinforcement has been shown to be an effective way to motivate a child to wear an orthodontic device. Once the desired behavior has been established, the reinforcement can be faded out gradually.
lie6674X_06_c06_201-230.indd 210 3/14/12 4:38 PM
CHAPTER 6Section 6.2 The Problem of Maintaining Behavior
Figure 6.2: Effects of different reinforcement techniques
Effects of three different reinforcement techniques: social reinforcement in the form of praise, money with a delayed payoff, and money with an immediate payoff. The reinforced behavior was the wearing of an orthdontic device. Source: From Hall, R. V., Axelrod, S., Tyler, L., Grief, E., Jones, F. C., & Robertson, R. (1972). Modification of behavior problems in the home with a parent as observer and experimenter. Journal of Applied Behavior Analysis, 5, 53–64. Reprinted by permission of John Wiley & Sons, Inc.
T im
e O
rt h
o d
o n
ti c
D e
v ic
e W
o rn
( p
e rc
e n
t)
Days
10 20 30 40 50 60 70 0
0
100
80
60
40
20
baseline 1
social reinforcement
delayed monetary
immediate monetary
baseline 2
immediate monetary
postchecks
At this point, the reinforcement program was discontinued, and the amount of time Jerry spent wearing the device immediately declined to 64%. This still represented a substantial improvement over the original figure of 25%, but the change was not sufficient to cure Jerry’s dental problems. The immediate reinforcement condition was therefore reinsti- tuted, and Jerry returned to wearing the device reliably (the percentage this time was 99%). Instead of terminating the program abruptly, his mother now faded it out gradu- ally. The frequency with which Jerry’s behavior was checked was gradually reduced from five times a day to only once every two weeks, and the behavior was now maintained. Eight months later, Jerry’s dentist told him that he had made great progress and no lon- ger needed to wear the device. By using reinforcement and then fading it out gradually, Jerry’s mother was able to establish and maintain a behavior that years of scolding and nagging had proved powerless to influence.
lie6674X_06_c06_201-230.indd 211 6/5/13 11:38 AM
CHAPTER 6Section 6.3 Harmful Effects of Reinforcement
6.3 Harmful Effects of Reinforcement The material reviewed in the previous sections testifies to the beneficial effects that rein- forcement can have when used properly. When encouraged to use reinforcement more fre- quently, however, parents and teachers sometimes react with suspicion, especially when the proposed reinforcer is a material one such as tokens or money. In this section, we will consider some of the reasons for this suspicion and the extent to which it might be justified.
Bribery One common objection to the use of reinforcement is that it seems to be a form of brib- ery. Why should a child be offered money or other rewards to mow the lawn or do other chores? Many people perceive these tasks as a duty; if so, a material inducement is noth- ing more than a bribe.
This view has some appeal: There is something disturbing about offering a reward to get someone to do something they should be doing anyway. However, we need to consider this issue in the context of available alternatives. There is no problem if children accept responsi- bility, but what alternatives are available if they do not? We could admonish them to do their duty, or threaten them with punishment if they disobey, and these strategies might be appro- priate in some circumstances. In at least some situations, however, these disciplinary tech- niques are ineffective as well as unpleasant. Consider the examples we have already seen: Robbie’s teachers tried punishment to eliminate his misbehavior without success, and eight years of reprimands had no effect on getting Jerry to wear his orthodontic device. In both cases, however, the introduction of reinforcement led to a rapid and substantial improvement in behavior, which was then main- tained even after reinforcement was gradually discontinued. These examples do not prove that reinforcement is always preferable to punishment, but they do suggest that reinforce- ment can be more effective than traditional forms of discipline in at least some circumstances, and might avoid harmful side effects that sometimes come with pun- ishment (see Chapter 6). O’Leary, Poulos, and Devine (1972) dis- cuss other issues concerning the relationship between reinforce- ment and bribery.
Greed A second objection to the use of rewards—particularly material rewards—is that they pro- mote greed. If children were offered $20 for cleaning their rooms, in this view, they would
These five-year-old twins from the 1940s are completing chores on the family farm. It is doubtful that they had to be motivated to do their chores with any type of positive reinforcement. They were probably expected to contribute to the family business without complaint.
lie6674X_06_c06_201-230.indd 212 3/14/12 4:38 PM
CHAPTER 6Section 6.3 Harmful Effects of Reinforcement
soon begin demanding money for doing other chores too, rather than accepting the chores as a necessary aspect of cooperative living. In fact, we have already encountered indirect evidence that material reinforcers can have this effect. In Chapter 4, we saw that extended exposure to a particular reinforcer deval- ues lesser reinforcers: A rat that had previously received 256 food pellets for running down an alley would not run nearly as fast to obtain 16 pellets as a rat that received 16 pellets all along. In other words, it looked very much as if the 256-pellet rat had become greedy!
Similar effects have sometimes been observed in applications
involving humans and material reinforcers. In one study, delinquent female adolescents were reinforced with money when they behaved appropriately in class—for example, not talking while the teacher was explaining something. The program was run during classes held in the morning, and it produced a significant improvement in students’ behavior. However, disruptive behavior increased during the afternoons, when the program was not in effect. As one student said to the experimenters, “If you don’t pay us, we won’t shape up” (Meichenbaum, Bowers, & Ross, 1968, p. 349).
To avoid these kinds of problems, most reinforcement programs begin by using relatively mild reinforcers such as social praise. Material reinforcers are used only if these milder forms of intervention prove ineffective.
Undermining Intrinsic Motivation A further objection to the use of reinforcers is that they can devalue the activity on which they are contingent. One view is that a person should be directed toward a certain behav- ior by intrinsic motivation—motivation that comes from the activity itself rather than from any consequences that might follow it. In the words of A. S. Neill, a Scottish educator who founded an influential school known as Summerhill,
The danger in rewarding a child is not as extreme as that of punishing him, but the undermining of the child’s morale through the giving of rewards is more subtle. Rewards are superfluous and negative. To offer a prize for doing a deed is tantamount to declaring that the deed is not worth doing for its own sake. .. . A reward should, for the most part, be subjective: self- satisfaction for the work accomplished. (Neil, 1960, pp. 162–163)
Some parents use “reward charts” or “sticker charts” to reinforce their children’s good behavior and adherence to household rules, where the particular reward may range anywhere from food treats to stickers to toys. However, some believe that a milder reinforcer such as social praise should be used before resorting to a reinforcement program based on material rewards.
lie6674X_06_c06_201-230.indd 213 3/14/12 4:39 PM
CHAPTER 6Section 6.3 Harmful Effects of Reinforcement
In practice, it is sometimes difficult to distinguish precisely intrinsic and extrinsic reinforc- ers. Take eating, for example. Should eating be considered intrinsically motivated because the pleasure derives from eating itself, or is the food an extrinsic reinforcer? In theory, the distinction seems reasonably clear: Intrinsically motivated behaviors are those that are relatively independent of external or arbitrary reinforcers.
Support for Neill’s view comes from a study by Lepper, Greene, and Nisbett (1973). The purpose of their experiment was to investigate the effects of reinforcement on children’s behavior in draw- ing pictures. In the first phase, the spontaneous level of drawing was determined by providing a nursery class with free access to felt-tip markers and paper and observing how much time they spent drawing during a three-hour period. One week later, the children were told that there was a visitor who would like to see what kinds of pic- tures children draw with markers. A reward group was told that they would receive a Good Player award—consisting of a card with a gold star, a red ribbon, and their names inscribed—if they drew a picture. A control group was also asked to draw a picture, but no reward was mentioned.
To test the effects of the reward, markers were again made available in the nursery one to two weeks later. Children in the control group spent almost exactly the same amount of time drawing as they had during the baseline phase, but the children who had been rewarded spent only half as much time as they had before.
Determinants of Undermining If reinforcement reduces long-term interest in an activity, why was it so effective in the studies reviewed earlier in this chapter, in which changes in behavior were maintained even after reinforcement was discontinued? Clearly, reinforcement does not always reduce interest; the outcome must somehow depend on the particular circumstances in which it is used. In this section we will consider several possible factors suggested by recent research.
Intrinsic or Extrinsic?
One obvious difference between the Lepper group study and earlier applications is that the activities reinforced in most of the earlier studies were not all that exciting to begin with. Robbie, for example, hardly derived pleasure from studying, nor did Jerry enjoy wearing his orthodontic device. In contrast, the Lepper study involved reinforcing a very attractive activity—drawing pictures. Perhaps reinforcement ultimately reduces interest only when intrinsic interest is high to begin with.
A. S. Neill (1883-1973) founded the progressive Summerhill School in Suffolk, England. The school is based on the philosophy that children learn best with freedom from coercion.
lie6674X_06_c06_201-230.indd 214 3/14/12 4:39 PM
CHAPTER 6Section 6.3 Harmful Effects of Reinforcement
Support for this hypothesis comes from Lepper, Greene, and Nisbett’s data. During the base- line phase, most of the children spent considerable time draw- ing; for these children, as we have seen, reinforcement sig- nificantly reduced interest. Some children, however, showed little interest in drawing initially, and these children became more interested in drawing following the reward. In practical terms, if a child hates lawn mowing, there is probably little danger that his or her interest will be reduced by the offer of a reinforcer; in fact, the pleasure derived from earn- ing money and feeling grown-up might actually enhance interest. However, for those children who already enjoy mowing lawns—a rare and much prized species—the offer of a reward might be more likely to prove counterproductive.
Coercion
Lepper (1981) and Deci and Ryan (1980) have suggested that when children are reinforced for engaging in an activity, they might feel that they are being controlled or manipulated. This sense of being controlled is aversive and could be responsible for their subsequent loss of interest in the task. According to this analysis, whether reinforcement will have a damaging effect should depend on whether recipients perceive it as an attempt to control their behavior. To test this hypothesis, Ryan (1982) gave students a number of interest- ing puzzles to solve. In one group, the experimenter said “Good” whenever the students solved a puzzle; in a second group, the experimenter said “Good, you’re doing as you should.” As predicted by the control hypothesis, subsequent interest in the task was sig- nificantly lower in the second group (see also Feehan & Enzle, 1991). The more we feel controlled, the less we enjoy the task we are being asked to perform.
Competence
We might expect quite different results if reinforcers were delivered in a way that encour- aged feelings of competence. If reinforcement was contingent on the quality of perfor- mance rather than simply on completing the task, it might be more likely to increase feelings of competence and thus lead to greater enjoyment and interest. In one study sup- porting this prediction, Enzle and Ross (1978) offered university students $1.50 for work- ing on difficult puzzles. In one group, the reward was promised simply for participating in the experiment; in a second group, it was contingent on achieving a level of competence well above average (this level was not specified, but all the subjects in this group were told that they had attained this high level). As in earlier studies in this area, subjects who
Finger painting is an intrinsically pleasurable activity for young children, who will engage in it without much encouragement. In some circumstances, providing a reward for such activities— especially a material reward such as money or a toy—can reduce interest rather than increase it.
lie6674X_06_c06_201-230.indd 215 3/14/12 4:39 PM
CHAPTER 6Section 6.3 Harmful Effects of Reinforcement
were reinforced simply for par- ticipation showed significantly less interest in the task after the experiment was over, but sub- jects reinforced for their skill showed greater interest. When reinforcement implies a greater level of competence, the plea- sure we experience is likely to enhance our enjoyment of the task rather than diminish it: “This is something I’m good at; what fun!” (See also Cameron, Pierce, Banko, & Gear, 2005.)
Evaluation The discovery that rewards can undermine interest in a task led to a flurry of research, and no
little controversy, with behaviorally-oriented psychologists tending to emphasize the advantages of reinforcement and cognitive psychologists the drawbacks. (See, for exam- ple, Cameron, Banko, & Pierce, 2001; and Henderlong & Lepper, 2002.) As research accu- mulated, however, the differences narrowed, and there is now broad agreement over the circumstances in which undermining is more likely.
In summarizing this consensus, the first point to note (and emphasize) is that rewards are generally used to encourage behaviors that people don’t enjoy, not those that they do. Children, for example, are usually reinforced for engaging in activities they find unattractive—eating spinach, cleaning their rooms, doing homework—and in these circumstances there is little danger of undermining interest. The risk is greater when behavior is intrinsically motivated, but even here praise is likely to encourage interest if it leaves a child feeling competent and valued (Cameron & Pierce, 1996; Henderlong & Lepper, 2002). To use a concrete example, suppose that you wanted to encourage chil- dren to practice the piano. It would probably be better to praise them for practicing rather than to offer a material reward, because praise seems less likely to be perceived as a mechanism of control. Moreover, this praise would be more likely to encourage long-term interest if it emphasized their competence (“That sounds lovely; you’ve really improved”) rather than their obedience (“That’s wonderful; you’ve practiced for an hour just as you were supposed to”).
The Principle of Minimal Force
Despite the sometimes spectacular success of applied reinforcement programs, parents and teachers often resist the use of reinforcement, and it is now clear that at least some of their concerns are justified. Rewards can sometimes encourage greed or lead to a sense of being controlled, which may reduce long-term interest in the reinforced activ- ity. This doesn’t mean, however, that we should never use reinforcement. When a task
A famous bicyclist such as Lance Armstrong, who receives cash prizes and prestige every time he wins a race—in effect, being rewarded for a high level of competence—will have greater levels of enjoyment in the sport than a bicyclist who is rewarded simply for his participation in a race.
lie6674X_06_c06_201-230.indd 216 3/14/12 4:39 PM
CHAPTER 6Section 6.3 Harmful Effects of Reinforcement
is unattractive, reinforcement can be a far more pleasant—and effective—technique than are alternatives such as threats or admonitions to be good. To minimize the problem of harmful side effects, current evidence suggests that when reinforcement is used, it is best to follow what might be called the principle of minimal force—that is, to use the least powerful reinforcer that is likely to be effective (see Lepper, 1981). In general, it is best to start with relatively mild reinforcers such as praise, turning to material reinforcers only if praise proves ineffective. Whatever the reinforcer chosen, wherever possible it should be administered in a way that encourages feelings of competence rather than mere obedience.
Promoting Autonomy
No one likes to be manipulated or controlled, and the more reinforce- ment is perceived as part of a caring or supportive relationship, the more likely it is to be effective. One way to encourage this feeling of coopera- tion is to involve the potential recip- ient of a reward in the design of the program—having the recipient help choose the goals, the reinforcers to be used, and the contingencies. An interesting example comes from a study by Ludwig and Geller (1997). Although the study did not involve reinforcement, it nicely illustrates the importance of individuals par- ticipating in the design of programs affecting their behavior. The study focused on a pizza firm’s desire to reduce accidents involving their drivers. One group of pizza deliverers were assembled for a meeting to discuss the impor- tance of coming to a full stop before joining the main road outside their store; they decided for themselves what targets to set for the percentage of occasions on which drivers should behave in this way, and they then received feedback for several weeks on the group’s suc- cess. A group at another store was treated similarly, except that their managers specified the targets. (The targets were set to match that chosen by the first group.)
The drivers’ behavior was monitored without their knowledge, not only during the four weeks of the program but for 5½ months after it ended. Both groups showed similar behavior in meeting the explicit target of coming to a full stop, but drivers who had par- ticipated in the target setting were found to also improve in other safety behaviors not mentioned in the program. Figure 6.3 shows the percentage of occasions on which driv- ers signalled before turning onto the main road. The subjects in the participatory group improved substantially on this behavior, and the improvement was maintained even after the program was terminated. The performance of subjects who had been assigned targets, however, deteriorated after the program was terminated, and was indistinguishable from that of control subjects who were never asked to change their safety behavior.
Research suggests that pizza drivers are more likely to stop dangerous driving practices if they participate in setting targets rather than having them imposed. None of us— workers, children, students—like to feel that others are trying to control our behavior.
lie6674X_06_c06_201-230.indd 217 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
Figure 6.3: Percentage of turn signal use by pizza-delivery drivers
Percentage of turn signal use by pizza delivery drivers who participated in the setting of goals or who had the goals assigned. Those who were allowed to set their own goals were more likely to change their behavior. Source: Adapted from Ludwig & Gellar, 1997
P er
ce n
t T u
rn S
ig n
al U
se
Experimental Conditions
70
60
50
40
30
20
90
80
Baseline Intervention Follow-up
participatory
assigned
All attempts to change others’ behavior can potentially be seen as coercive, and much depends on exactly how these programs are implemented. The more individuals feel that they are valued, and their needs and wishes are being considered, the more likely it is that they will cooperate in changing their behavior over the long term. (See also Grolnick & Ryan, 1989.)
6.4 Self-Control We have seen that the greatest weakness of reinforcement programs lies not in estab- lishing behaviors initially but, rather, in maintaining them when the program is dis- continued. If no reinforcement is provided, the response might simply extinguish. Also, in cases where the reinforcement program is seen as coercive or manipulative, interest
lie6674X_06_c06_201-230.indd 218 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
in the task can actually be lower than it was originally. A potential solution to both of these problems is to encourage self-control—that is, to train people to control their own behavior rather than rely on reinforcement from external sources. Before considering how we might help people exercise greater self-control, though, we need to discuss what we mean by this term.
The Concept of Self-Control Let us start with a concrete example. Suppose that a man named Tom wants to lose weight, but every time he tries to diet he fails. He has a particular weakness for chocolate bars and eats several every night before going to bed. Each time he diets he vows that he will give up these chocolates, but when he goes to bed he just can’t resist eating them.
Willpower
How can we explain Tom’s inability to diet? The conventional explanation for failures like this is poor willpower—Tom just doesn’t have the self-control or willpower to make himself adhere to his diet. But what does it really mean to blame his failure on lack of will- power? The term willpower implies that we have a will that we can use to make ourselves do what we want, but is there really one part of our mind that forces other parts to obey its bidding? If so, and if some people have stronger wills than others, we should expect that individuals with strong wills would be uniformly good at making themselves perform difficult tasks. This, however, does not appear to be the case. For example, you might
know people who are very good at making themselves study, but cannot resist cigarettes or overeating (see also Mischel & Mischel, 1977). Another prob- lem with the concept of will- power is that it seems to leave us helpless. If some people have greater willpower than others, what can people like Tom, who are deficient, do to suddenly endow themselves with more?
Difficulties such as this have persuaded some psychologists that the concept of willpower is not really useful in explain- ing self-control—indeed, that
it is simply an explanatory fiction that we invoke to explain behaviors that we don’t understand. If Tom has difficulty dieting, we attribute his difficulty to poor willpower, but we have no independent evidence that willpower exists—we can’t see it, and Tom can’t feel it. The explanation is circular: We attribute Tom’s failure at dieting to poor willpower, but the only way we know that he lacks willpower is that he is having dif- ficulty in dieting. It is a bit like the medieval belief that people who behaved strangely were possessed by demons—the strange behavior was attributed to demons, but the
Some individuals are more motivated than others to work out at the gym, but this isn’t necessarily a matter of willpower.
lie6674X_06_c06_201-230.indd 219 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
only evidence for the existence of demons was the strange behavior. Both explanations might make us feel better because they seem to provide an explanation for behavior that would otherwise be mysterious, but they are really only pushing the mystery a step fur- ther away. In the case of willpower, lack of willpower seems to explain Tom’s difficulty in dieting, but we don’t then consider what willpower really is, and why Tom has less of it than others have.
The claim that willpower doesn’t exist might strike you as obviously mistaken because we have all been exposed to this concept for so long that we simply take it for granted. Suppose for the moment, though, that the claim were correct and willpower really did not exist—how then could we explain why some people succeed at dieting or giving up smoking, but others fail?
Reinforcement Contingencies
Behavioral psychologists such as B. F. Skinner (1953) and Howard Rachlin (1974) have proposed one possible explanation. They argue that difficulties in self-control arise not from a lack of willpower but, rather, from reinforcement contingencies that favor imme- diate gratification over long-term interests. Consider again Tom’s problems with eating chocolates. If he eats a chocolate bar before going to bed, he obtains immediate reinforce- ment from its taste. If he leaves it uneaten, he will lose some weight, but the amount he loses will be so small as to be undetectable. Only if he diets for an extended period will he lose enough so that he can begin to see the difference in a mirror, or to feel healthier. In other words, although there are strong reinforcers available for dieting, they are substan- tially delayed. Given a choice between eating a chocolate bar and not eating it, Tom may choose to eat it because he obtains a small but immediate form of reinforcement for doing so; dieting produces greater reinforcement, but only after a much longer delay.
Snight: Reat S R
chocolate
Rdiet S R
weightloss
The situation is much like that faced by the pigeons in the Rachlin and Green experiment that were given a choice between a small amount of food immediately or a large amount after a delay, or students who have to choose between studying and going to a movie (see Chapter 4). And unfortunately for Tom, the outcome is also the same—the small but immediate reinforcer exerts greater control.
In this view, choosing between eating chocolate or abstaining is no different from choos- ing whether to have a hamburger or a hot dog for lunch, or what clothing to wear to a party. They are all simply choice situations in which we choose between alternative responses, and the choice we make largely depends on the reinforcement available for each. As to why some people are better at refusing chocolate, one factor might be differ- ences in the reinforcement contingencies affecting them. If Tom faces greater-than-average stress, for example, then the soothing properties of chocolate may make it a more power- ful reinforcer for Tom than for others, thereby increasing the likelihood of his eating it. Also, people who are good at dieting might have learned coping or self-control responses that help them in these situations.
lie6674X_06_c06_201-230.indd 220 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
A Painful Example
To illustrate the concept of self-control responses, we will use an experiment by Kanfer and Seidner (1973). To measure self-control, they asked one group of subjects to keep one hand in a bucket of ice water for as long as they could stand it. The water was very cold, and the average immersion time was only 57 seconds. A second group was given access to a slide projector containing pictures of holiday scenes they could look at while their hands were in the water. The subjects in this group were able to keep their hands immersed for an average of 149 seconds, almost three times as long as those in the first group. Note that this result cannot be explained by willpower—because subjects were assigned to groups at random, and levels of will- power in the two groups would have been roughly equal. Sub- jects in the slide group could keep their hands in the water because they had a response they could perform—looking at pictures—that distracted them from the pain.
In this analysis, self-control is viewed simply as a set of responses that individuals can perform to alter their own behavior. This might at first seem contrary to the principle of determinism that we discussed in Chapter 1: If all behavior is determined, you might wonder, how can people be said to con- trol their own behavior? Skinner (1953) suggested a solution to this apparent paradox. His argument was that behavior is indeed controlled by the environment, but that an individual’s behavior can also alter that environment. An individual can thus perform a response now to alter his or her environ- ment and thereby indirectly alter the probability of his or her future behavior. In the Kan- fer and Seidner experiment, for example, when subjects turned on the slide projector they changed their visual environment, and this helped them reduce the amount of attention they paid to their pain. The greater self-control of individuals in the slide group was thus not because of greater willpower but, rather, because they used a specific response that allowed them to modify the situation.
Self-Control Techniques In considering how you can change your behavior, one useful principle that we have already encountered is that of stimulus control. We saw in Chapter 5 that when a response is reinforced, this does not usually result in a general increase in the probability of the response but rather in an increase in the specific situation in which the response was reinforced. The idea is that the stimuli present when a response is reinforced become
If you were offered a gorgeous chocolate dessert while trying to diet, one way to resist would be to distract yourself, perhaps by imagining a vacation on a palm-fringed, white sand beach, lapped by turquoise waters.
lie6674X_06_c06_201-230.indd 221 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
associated with the response, so that the response is more likely to occur when these stimuli are present.
This principle turns out to have several useful applications in programs for changing behavior. One example we have already encountered in Chapter 4 was Bootzin’s work on insomnia. To help those who had difficulty falling asleep at night, Bootzin (1972) advised his clients to get up from bed whenever they had difficulty sleeping, so that the stimulus of being in bed would not become associated with their restless behavior and thereby come to elicit it. This treatment, as we have seen, proved very effective.
Another interesting application of stimulus control was reported by Stuart (1967), as part of a program he developed for treating obesity. As one part of his program, Stuart asked his patients not to eat while engaging in other activities such as reading or watching tele- vision. The purpose was to break the association between these stimuli and eating, so that there would be fewer situations that elicited this behavior. This program proved to be remarkably effective—his eight patients lost an average of 38 pounds in one year, mak-
ing it one of the more successful dieting programs ever reported. Subsequent studies using his techniques have largely con- firmed this success, but, as in most other diets, participants often find it hard to maintain their weight loss after the pro- gram has ended (Wadden, Fos- ter, & Letizia, 1994). Because of the very powerful reinforcement that food provides, it is not easy to change eating behavior, but the principle of stimulus con- trol does seem to help and can contribute to impressive weight losses over periods of at least a year. (See also Levy et al., 2007.)
Self-Reinforcement
We have suggested that one of the main reasons that reinforcement is sometimes inef- fective is the delay between response and reinforcer. Consider the behavior of study- ing. There are a number of powerful reinforcers for studying—good grades, parental approval, improved career prospects, and so on—but these reinforcers are delayed for weeks, months, or even years. To take a wildly hypothetical example, imagine a college student who has to choose between reading a psychology text and going out on a date. The reinforcement for the date is relatively immediate; the reinforcement for studying is delayed days or weeks. From a reinforcement perspective, it is hardly surprising that many students have difficulty studying under these conditions.
When the environment does not provide immediate reinforcement for a behavior, one possible strategy is for individuals to reinforce themselves. If you wanted to increase the
Eating dinner in front of the TV is common in some families, but research shows that an association between TV watching and eating that develops from this behavior may contribute to obesity.
lie6674X_06_c06_201-230.indd 222 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
amount of time you spent studying, for example, you could reinforce your own studying behav- ior by, say, allowing yourself a 15-minute break whenever you completed 20 pages of a text. But would this technique really work? If you found it difficult to study, would you really wait until you had completed a difficult assignment before taking a break, or would you just give up and take the break anyway?
The answer, according to Skinner, would depend on your past history. As we have seen, Skinner viewed self-control simply as a set of responses, with the same properties as other responses. In particular, he argued that self-control responses are learned in the same way as other behaviors, through a combination of reinforcement and pun- ishment. Whether children will learn to reinforce themselves—for example, with candy or, more likely with praise (“I’ve been such a good girl”)— will depend on whether such behavior is in turn reinforced by others. When children praise them- selves appropriately, others might in turn praise them. (Tommy: “I did a good job cleaning my room, didn’t I, Mommy?” Mother: “Yes, Tommy, you did it beautifully.”) If they cheat, on the other hand, they might be reprimanded. Provided that we get enough appropriate feedback, we eventually learn to praise ourselves only when such praise is merited, and this self-praise can then help to maintain our behavior.
Developing Self-Control To illustrate how a behavior such as self-reinforcement might be learned, we will use an example reported by Drabman, Spitalnik, and O’Leary (1973). The subjects were 10-year- old boys in a class for children with academic and emotional problems. Eight of the most disruptive boys in the class were selected for special training, and a token economy was established in which the boys were given points on a five-point scale for good behavior and for completing assignments; at the end of each lesson, the points could be exchanged for cakes, candies, or pennies.
The program was highly effective: The frequency of disruptive behavior fell by two-thirds, and the average number of assignments completed rose from 83 to 130. It was not pos- sible, however, to maintain the token economy indefinitely. What, then, could be done to ensure that the gains would be sustained once the program was withdrawn?
Because a teacher could not always be available to provide reinforcement, Drabman and his colleagues decided to train the children to reinforce themselves. At the end of each les- son, the boys were to award themselves points on the basis of how they had behaved, with these points then being exchanged for other reinforcers in the usual way. To ensure that
College students living in dormitories for the first time may find it difficult to make appropriate choices between studying and engaging in social activities that are more immediately reinforced.
lie6674X_06_c06_201-230.indd 223 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
the boys would reinforce themselves appropriately, the researchers instituted a training program in which the teacher initially reinforced the boys for accurate self-reinforcement. Once this behavior had been learned, the frequency of checking by the teacher was pro- gressively reduced.
Specifically, the teacher monitored the boys’ behavior during the self-reinforcement phase, and at the end of each lesson the boys’ self-ratings were compared with the teacher’s rat- ings. If the boys’ ratings were within one point of the teacher’s, they received the points they had given themselves; if the ratings matched exactly, the boys received a bonus point; but if they had deviated from the teacher’s ratings by more than a point, they received no points. To reduce the likelihood that the checking would be seen as a form of control, the experimenters explained that being selected for checking was a privilege because only those boys who were checked would have the chance to earn bonus points.
Over days, the proportion of boys selected for checking was gradually reduced (fading), until all the boys were receiving whatever points they had awarded themselves with- out any formal checking. On days when their self-ratings exactly matched the teacher’s, though, the teacher strongly praised them. The results are shown in Figure 6.4. During the final phase, in which the boys received whatever points they had awarded themselves, their behavior was not only maintained at the levels achieved when the teacher controlled reinforcement, but, if anything, actually exceeded those levels.
lie6674X_06_c06_201-230.indd 224 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
Figure 6.4: Mean number of disruptive behaviors in successive phases of a self-control program
Even when the students began practicing self-reinforcement rather than being reinforced by a teacher, their levels of disruptive behavior remained low. Source: Adapted from Drabman, Spitalnik, & O’Leary, 1973
M ea
n N
u m
b er
o f D
is ru
p ti
ve B
eh av
io rs
Phases of Reinforcement
70
60
50
40
30
20
10
0
90
80
to ke
n e
co n
o m
y
m at
ch in
g
fa d
in g
1
fa d
in g
2
fa d
in g
3
fa d
in g
4
se lf-
ev al
u at
io n
b as
el in
e
Returning to our studying example, the reason that some students spend more time study- ing than others could be, in part, because they obtain immediate secondary reinforcement when they complete an assignment, in the form of a feeling of pride. And the source of this pride could be that they were differentially reinforced when they were younger for sticking to commitments—praised when they did, but not praised when they didn’t. This analysis is speculative, but the results of the Drabman et al. study suggest that children can learn to reinforce their own behavior if the appropriate use of self-reinforcement is itself reinforced (see also Wood & Flynn, 1978).
lie6674X_06_c06_201-230.indd 225 3/14/12 4:39 PM
CHAPTER 6Section 6.4 Self-Control
More broadly, it appears that self-control procedures can be very powerful if they are backed-up by occasional monitoring and support from others (for example, Ward & Carnes, 2002). In the case of obesity, for example, self-control procedures have been very effective in helping people lose weight, but participants usually regain this weight once treatment is terminated. To counter this problem, Latner et al. (2002) tried incorporating external sources of reward and punishment into a self-control program. For example, par- ticipants were told that they would be dropped from the program if they did not adhere to the regulations during the first few months. This combination of self-control methods with external consequences has produced the most impressive results to date, with partic- ipants maintaining losses of more than 15% of their body weight for more than five years.
Improving Your Studying To summarize some of the self-control principles we have been discussing, we will conclude by briefly considering how you can use self-control principles to increase the amount of time you spend studying. According to several successful programs (for example, Goldiamond, 1965; Fox, 1966), your first step should be to find a quiet spot where you can work with minimal disturbance—for example, an isolated desk in a library. Begin by setting yourself a modest target for how long you will study, and then reinforce yourself (self-reinforcement) when you reach your target (the reinforcer could be coffee, a break with friends, or even just a notation on a special record card—an accurate record of progress can be a surprisingly powerful reinforcer). Then, over days, gradually increase your target so that the amount of time you spend studying before you reinforce yourself is gradually increased (shaping).
For this strategy to work, it is important to choose an effective reinforcer. Some evidence suggests that a public declaration of both your goals and your progress—for example, posting a graph of your studying time where your friends can see it and thus encourage you—can also be important (see Hayes et al., 1985). Also, you need to set goals that are realistically attainable. If you have difficulty concentrating, you might need to set your initial goal at only 15 minutes, or even 5 minutes, and then increase your target gradually.
To maximize the probability that your desk will become a cue for studying and not for other behaviors, ensure that studying is the only activity you engage in while there. If you feel an uncontrollable urge to daydream or have a snack, leave immediately, and return only when you feel able to resume concentrating on your work (stimulus control).
These are by no means the only useful techniques for improving studying habits. (For some other techniques, see Fox, 1966; Weinstein & Meyer, 1986.) However, if you do want to improve your studying, you might find this approach a helpful component of a broader program based on the following:
• Practice. Alas, repetition matters: The more you study, the more likely you are to remember. However, how much you remember depends not just on the amount of time you spend practicing but also on the spacing of this practice. Your teach- ers were right: Cramming the night before an exam may be better than not study- ing at all, but not nearly as effective as studying spread over a semester.
• Thinking. The better you understand material, the better you will remember it. Instead of just reading material in a rote fashion, straight through at a gallop, think as you read—for example, try to relate the material to your own experience.
lie6674X_06_c06_201-230.indd 226 3/14/12 4:39 PM
CHAPTER 6Review Questions
• Practicing retrieval. Recent research has shown that time spent recalling or retriev- ing material can be as effective as, or even more effective than, time spent in rereading it. After you finish a chapter or a section, try closing your textbook and seeing how much you can remember without looking at it. The more you do this—for example, the next day while walking to a class—the more likely you are to recall the material on an exam.
Summary and Review
• When reinforcement principles are used properly, they can be surprisingly effec- tive. Examples discussed include reducing classroom misbehavior, improving homework, and treating autism.
• When reinforcement is discontinued, the behavior that was rewarded may extin- guish. To provide more time for reinforcers in one’s natural environment to take over, useful techniques include using partial reinforcement in training, reinforcing behavior in a variety of settings, and fading out the use of reinforcement gradually.
• Reinforcement can sometimes have harmful side effects. If material rewards such as money are used, there is a danger of encouraging greed. Also, when children are reinforced for activities that they already enjoy, this can paradoxically under- mine interest, particularly if the child feels coerced.
• To minimize side effects, the “principle of minimal force” should always be applied: Start by using the mildest reinforcer that is likely to be effective (in most cases, praise).
• One way to reduce people’s feelings of being controlled by others is to encour- age self-control. According to Skinner, self-control is simply a set of learned behaviors, acquired because they change the probability of future behaviors and thereby shorten the time to reinforcement. Self-control in this view is not a matter of willpower but of learning effective strategies such as stimulus control and self- reinforcement. These techniques have proven very effective in helping people to lose weight and increase the time they spend studying.
Review Questions
1. The “principle of minimal force” suggests starting with relatively mild reinforc- ers, such as praise, whenever possible. Can social reinforcers such as praise really modify difficult behaviors? If they fail, what other reinforcers can be used?
2. What can be done to increase the likelihood that behaviors will persist long enough after a reinforcement program is terminated to allow natural reinforcers to acquire control?
3. What are the potentially harmful effects of reinforcement? In what situations are these most likely to occur?
4. What are some of the techniques by which people control their own behavior? How can they be applied to studying?
5. How does Skinner explain self-control? How does his account differ from that of willpower? What does each approach say about variability in self-control? In other words, should we expect a person who shows strong self-control in one situation (for example, giving up smoking) to also have above-average self-con- trol in other situations (for example, studying)?
lie6674X_06_c06_201-230.indd 227 3/14/12 4:39 PM
CHAPTER 6Concept Check
Concept Check
1. What does the term behavior modification refer to?
a. The application of operant conditioning principles in an attempt to change human behavior
b. A gradual reduction in how often behavior is reinforced c. The learning of a new behavior through observation of the behavior of others d. A reduction in the intensity or frequency of reinforcement
2. What was so impressive about the results of Lovaas’ studies on the treatment of autistic children?
a. Children treated in the experiment demonstrated little improvement in IQ but remarkable improvement in social skills as compared to children in the con- trol group.
b. Children treated in the experiment demonstrated little improvement in social skills but remarkable improvement in IQ as compared to children in the con- trol group.
c. Children treated in the experiment demonstrated remarkable improvement in both behavior and IQ as compared to children in the control group.
d. Children treated in the experiment demonstrated remarkable improvement in analytical skills as compared to children in the control group.
3. Material reinforcement has been criticized as a way of negating the ability of a person to be directed toward
a. bribery. b. coercion. c. intrinsic motivation. d. self-punishment.
4. The more people feel that they are ________, and that their needs and wishes are being considered, the more likely it is that they will __________in changing their behavior over the long term.
a. coerced, participate b. disbelieved, participate c. mistreated, cooperate d. valued, cooperate
5. When the environment does not provide immediate reinforcement for a behavior, what is a possible strategy for individuals to use instead?
a. Determine that the behavior modification was not complete and try again until the environment provides the correct reinforcement
b. Reinforce themselves c. Seek out immediate gratification even if it goes against the desired changed
behavior d. Seek reinforcement from others
Answers: 1) a, 2) c, 3) c, 4) d, 5) b
lie6674X_06_c06_201-230.indd 228 3/14/12 4:39 PM
CHAPTER 6Key Terms
Key Terms
ABA design In most experiments where the goal is to compare the effects of two conditions, A and B (e.g., reinforcement or no reinforcement), separate groups are used. In an ABA design, the same subject receives both conditions—for example, A, then B, then A again, and so on.
behavior modification The systematic use of operant conditioning techniques such as reinforcement and extinction to change behavior.
extrinsic reinforcer Extrinsic means “exter- nal”; when the reinforcer for performance on a task is not inherent simply in perfor- mance of the task but rather follows its com- pletion, the reinforcer is said to be extrinsic. If you play a video game because you enjoy the game, this is an example of intrinsic reinforcement; if you play to earn praise or money, those would be extrinsic reinforcers.
fading A gradual reduction in the intensity or frequency of a stimulus. In reinforcement programs, this refers to a gradual reduction in how often a behav- ior is reinforced, in contrast to the abrupt termination of a program.
intrinsic motivation The motivation to perform a task that comes from the charac- teristics of the task itself rather than from any consequences that might follow (for example, praise or awards). If you play a game simply because you enjoy it, your motivation would be intrinsic to the task, or built in.
intrinsic reinforcer A reinforcer that arises naturally from the performance of a task, rather than one delivered subsequently by an external agent.
modeling The learning of a new behav- ior through observation of the behavior of others. The individual exhibiting the desired behavior is called a model, and observational learning is sometimes referred to as imitation (which emphasizes the behavior of the observer) rather than modeling (which emphasizes the behavior of the model).
partial reinforcement Any situation in which only some instances of a response are reinforced.
principle of minimal force The recom- mendation that where reinforcement or punishment is used, we adopt the least powerful consequence likely to be effective, in order to minimize the side effects that can occur with powerful consequences.
self-control For an individual to control his or her own behavior rather than relying on reinforcement from external sources.
token economy A systematic procedure for reinforcing behavior, in which tokens are made contingent on the performance of desired behaviors; the tokens can later be exchanged for backup reinforcers such as candy.
willpower The capacity to control one’s own behavior, emotions, and so on. Will- power is viewed as a limited resource, and some people are thought to have more of this capacity or resource than others.
lie6674X_06_c06_201-230.indd 229 3/14/12 4:39 PM
lie6674X_06_c06_201-230.indd 230 3/14/12 4:39 PM
