edpsy10natural

Artificial and Natural Reinforcement

Reinforcement can conveniently be divided into artificial (contrived) and natural (logical) reinforcement. This division applies to all three types of reinforcement. A reinforcer is artificial (contrived) to the extent that the recipient sees it as logically unrelated to the task which must be performed in order to obtain it. A reinforcer is natural (logical) to the extent that the recipient perceives it as logically related to the task he has to perform to obtain it. Few reinforcers are purely natural or artificial. Rather, reinforcers can be graded along a continuum, with some reinforcers being more natural than others.⁶ For example, reading a book because it is a good book is at the natural end of the curriculum. Reading a book when I hate the book but my parents will pay me five dollars for reading it is at the artificial end. Reading a book when I would rather watch television because I know my parents would be proud of me would be somewhere in the middle of the continuum. Likewise, reading a difficult book because I want to become a photographer and I know I can get a better job as a photographer if I get a college degree is somewhere in the middle.

The manner in which a reinforcer is introduced can influence the degree to which it is natural or artificial. For example, giving a child a dollar for getting an A on a spelling test sounds highly artificial. On the other hand, spending the dollar to celebrate a significant achievement sounds much less artificial.

All three types of reinforcement can be either artificial or natural.

As a general rule, natural reinforcers are preferred to artificial reinforcers. With natural reinforcers, the strengthening of the behavior does not rely on the availability of an outside agent to dispense the reinforcement. The contingent pleasant situation arises from the behavior itself. There is evidence that the unnecessary use of artificial reinforcers can actually reduce a learner's motivation to engage in the reinforced activities (Deci, 1975; Deci & Ryan, 1992; Lepper & Greene, 1978). In addition, educational environments that foster greater autonomy and self-determination are more likely to lead to higher levels of educational activity and achievement (Rigby et al, 1992). The underlying principle seems to be that if we have to bribe someone to play, then that person is likely to start viewing that play as work. This negative impact of artificial reinforcement is most likely to be a problem when the reinforcement is viewed as completely artificial and arbitrary - that is, when there is no reason to engage in the behavior other than to gain the reinforcer. More detailed discussions of the possible ill effects of artificial manipulation of children's behavior can be found in Glasser (1985) (who focuses on the learners' need to control their lives) and Kamii, Clark, & Dominick (1994) (who employ a constructivist perspective to focus on the learners' need to develop autonomy).

To the extent that supplementary reinforcers are based on principles of intrinsic motivation (discussed in Chapter 5), they are not as likely to have this negative impact (Lepper & Hodell, 1989). In addition, the manner in which external rewards are administered (for example, the degree to which they are coercive) also influences the likelihood that they will have a negative impact on motivation and learning (Ryan, Mims, & Kostner, 1983; Deci & Ryan, 1992). An important research finding is that artificial rewards may be important in enhancing interest among students who are not initially motivated, but there is no reliable data to describe ways in which extrinsic rewards can further enhance motivation among students who are already intrinsically motivated (Lepper & Hodell, 1989; Bandura & Schunk, 1981).

BOX 10.1
Why Do Teachers So Often Make Reinforcement Artificial?

It has always seemed astonishing to me how often students are practically forced to rely on artificial reinforcement. As an English teacher, I used to teach my students about the "structural elements" of short stories. My students read some really good stories and seemed to enjoy them; but most of them saw little sense in analyzing the tone, setting, theme, and character development in the stories. They learned about these elements because they believed that this would be necessary to get into college and because I worked hard to make this information interesting to them. This was artificial reinforcement. In retrospect, it would have made a lot more sense if I could have shown them that they could enjoy the stories even more if they would have understood and applied the concepts of tone, setting, theme, and character development while reading the stories. This would have been natural reinforcement.

When I myself was in a college math class, I recall raising my hand and asking the professor if there was any purpose in learning how to graph equations if we didn't plan to become engineers or math teachers. He replied with a simple "No." I got an A in that course, but I had to rely on artificial reinforcement. As I progressed to graduate school, I discovered that graphing equations had a lot to do with understanding statistical concepts and projections of various types. I had learned to hate those stupid equations because of the artificial reinforcement, and so I had an artificial animosity toward statistics that interfered with my learning. That teacher should have been able to find some way to convey to me the naturally reinforcing reasons for learning to graph equations. He could have appealed to my cognitive curiosity, to my fantasy, or to my sense of competition to make me enjoy devising the equations. This would have been Type I Reinforcement. As it was, I received extremely artificial Type III Reinforcement by avoiding the loss of my A if I managed to learn enough about those equations.

Box 10.2
Most Reinforcers Are a Combination of Natural and Artificial
The ambivalent value of artificial reinforcers is not really all that unusual. It may be useful to approach this concept from a different perspective. Purely natural reinforcement is actually rare; and people who rely on purely natural reinforcement in order to become motivated are likely to be unproductive. Purely natural reinforcement occurs only when the person receiving it would continue doing that behavior even if all other consequences for that behavior were withdrawn. For example, my son used to play on his high school's tennis team. Let's examine the extent to which he received natural reinforcement for this.

He played competitive matches for his school against other teams. (Would he have continued to do this if playing these games were the only consequence? He probably would have. We'll consider this to be natural reinforcement.)

He attended practices where he went through lots of drills. For example, he may have had to hit fifty shots in a row with his backhand stroke. (Would he have continued to do this if completing these drills were the only consequence? No; but he viewed them as important because he saw a direct connection between them and being successful in the matches against the other schools. These were less natural - but still fairly natural, as long as he saw their relevance.)

He had to run several miles a day during the off season to keep in shape for his tennis matches. (Would he have donr this if completing these runs themselves were the only consequence. No; and he objected to doing this, because running seemed to waste time and was exhausting. He 'was wrong about this; but he didn't see these runs as being closely related to success in tennis. So he viewed them as artificial.)

Occasionally, he had to go to meetings associated with tennis. The coach had to schedule these meetings in order to get forms filled out, to see that everyone knew what to do, etc. (Would he have gone to these meetings if they weren't necessary for being on the tennis team. No way. On the other hand, he bought into the logic that the paperwork was necessary, and so he viewed them as at least partly a natural result of wanting to play tennis.)

Now let's look at an academic example closer to your own life. Let's assume that there is a mythical educator who finds it to be intensely enjoyable to engage her students in active and productive learning. Let's examine the extent to which she might find taking an educational psychology course to be naturally reinforcing. She might have the following perceptions about the course:

During the course, she comes to an understanding of a concept that enables her to teach certain skills or topics much more effectively. (Would she continue to take this course if the prospect of frequent, additional similar insights were the only consequence? She poissibly might, assuming she really enjoyed teaching and planned to continue in that role.)

Occasionally she might come across concepts in the book that are difficult to understand. She clearly realizes that these concepts will eventually contribute to important insights, but they currently require more effort that she would like to spend on them. (Would she continue to work at this task if mastering those concepts were the only consequence of her continuing to study them? No. The value of these concepts lies in the close connection she sees between them and the insights she hopes to attain.)

In order to continue successfully in the course, she occasionally has to take tests and perform other exercises which she thinks are intellectually futile. (Would she continue to engage in these activities if succeeding at them were the only consequence? No. She may be wrong, but she would give these up if she had a chance. The only reason she is doing them is because she must do so to stay in the course.)

Registration. In order to get into the course, she has to stand in a line, talk to polite bureaucrats, and pay money. (Would she continue to do this if successful registration itself were the only consequence? Not unless she has a really lonely private life.)

These four perceptions have been listed in descending order with regard to natural reinforcement. Our mythical educator is likely to consider the overall impact of taking the course to be negative, unless the first perspective outweighs the other three. The positive impact of the course can be increased by making the first perspective even more positive or by minimizing the negative elements of the others (for example, by making registration a more painless experience).

Box 10.3
Quantifying the Pros and Cons of Reinforcement

Tables 10.1 and 10.2 attempt to clarify the positive and negative elements of artificial reinforcement by assigning numerical values to the various aspects of a single reinforcement context. Table 10.1 describes three students who were assigned to read Hamlet in their English class. All three received an A on the test because they read the play carefully.

Table 10.1. The reactions of three different students to studying hard and getting an A on a Hamlet test.

Event Internal Perceptions of Reward and Punishment Result

Bill gets an A on a test because he read Hamlet carefully.

R3 Bill feels good because he values the A.

P5 Bill feels bad because he dislikes the loss of control he experiences when he is coerced to do things against his will.

P8 Bill feels bad because he strongly disliked the play (because he could hardly understand it).

Bill is less likely to read things like Hamlet in the future. (P13 - that is P5 plus P8 - is greater than R3.)

Mary gets an A on a test because she read Hamlet carefully.

R3 Mary feels good because she values the A.

P5 Mary feels bad because she dislikes the loss of control hse experiences when she is coerced to do things against her will.

R8 Mary feels good because she thoroughly enjoyed reading the play.

Mary is more likely to read things like Hamlet in the future. (R11 - that is R3 plus R8 - is greater than P5.)

Joan gets an A on a test because she read Hamlet carefully.

R3 Joan feels good because she values the A.

P0 Joan is unconcerned about any loss of control. She thinks that it is perfectly sensible to follow the advice of an experienced teacher.

R8 Joan feels good because she thoroughly enjoyed reading the play.

Joan is much more likely to read things like Hamlet in the future. (R11 - that is R3 plus R8 - is much greater than P0.)

All three felt good because they valued the A. The R3 written next to that statement in the figure means that this had a reinforcing value of 3 (on a scale of 0 to 10) for each student. (All of these values are arbitrary. Different students would assign different values, depending on their previous experience, value systems, and other personality factors.)
Both Bill and Mary also felt bad because they have experienced a loss of control; and I have signified this by a value of P5, which means that they considered this aspect to have a value of 5 on a punishment scale ranging from 0 to 10. Joan, on the other hand, was unconcerned about any loss of control, and so I have signified this by a value of P0.
Bill hated the play (perhaps because he was thoroughly confused and frustrated while reading it); and I have signified this by a value of P8. On the other hand, both Mary and Joan enjoyed the play; and I have signified this by a value of R8.

Based on the previous values, we can expect the following:

Bill is less likely to read literature like Hamlet in the future. (The reinforcement he derived from Hamlet totaled 3 and the punishment totaled 13.)
Mary is more likely to read literature like Hamlet in the future. (The reinforcement she derived from Hamlet totaled 11, while the punishment totaled only 5.)
Joan is even more likely than Mary to read literature like Hamlet in the future, because her reinforcement totaled 11, while she experienced no punishment at all.

This analysis demonstrates the need to consider the individual cognitive backgrounds and personalities of individual students when considering the impact of reinforcement.

Box 10.4
Table 10.1 was an oversimplification.

Actually, there could be numerous aspects of reinforcement and punishment having an impact on a learner at one time. Table 10.2 diagrams a more complex possibility. This diagram shows seven factors that could influence Horatio, another student of Hamlet. The reinforcement and punishment values outside of parentheses suggest that Horatio would avoid reading literature like Hamlet in the future, because the overall punishing aspects outweigh the reinforcing aspects of reading Hamlet. The values in parentheses describe an alternate scenario, in which Horatio - with different reactions and feelings - would become more likely to read literature like Hamlet in the future.

TABLE 10.2 More complex reactions to studying hard enough to get an A on a Hamlet Test.

Event Reaction Result

Horatio gets an A on the Hamlet test.

Reaction on a Bad Day

Internal Perceptions of Reward and Punishment

Reaction on a Good Day

R3

He feels good, because he values the A.

R5

P5

He feels bad, because he doesn't like to be coerced to do things.

P1

P8

He feels bad, because he disliked the play.

P5

R2

He feels good, because his parents are proud of him.

R6

P8

He feels bad, because some of his friends think he is a nerd for getting an A.

P4

P5

He feels bad, because he feels pressured to get more A's

P1

R8

He feels good, because his grades may qualify him for a scholarship to attend college.

R8

The negative reactions (R=13; P=19) suggest that he will avoid things like Hamlet in the future.

The positive reactions (R=19; P=11) suggest that he will read things like Hamlet in the future.

NOTE: The Reaction on a Bad Day column (light green on left of the center column) suggests a possible set of relatively negative reactions by Horatio to the perceptions described in the middle column. (In other words, Horatio has a somewhat negative outlook on academic activities.) These negative reactions add up to a total value of Reinforcement = 13 and Punishment = 19; and under these circumstances Horatio would consider Hamlet to have been an overall negative experience.

The Reaction on a Good Day column (darker green on right of the center column) gives a set of more favorable reactions to the perceptions listed in the middle column. (In this second case, Horatio has a more favorable outlook on academic activities.) These more positive reactions add up to a total value of Reinforcement = 19 and Punishment = 11. Under these circumstances, Horatio would consider Hamlet to be an overall positive experience.

Artificial reinforcers are necessary when the natural reinforcer is initially not perceived as a pleasant set of circumstances available to the learner. This might occur, for example, if the learner does not currently know of the existence of the natural reinforcer or if the natural reinforcer cannot be attained until the distant future. In addition, sometimes a parent or teacher simply does not have control over the natural reinforcers which would influence the desired behavior. In such cases, it is often a good idea to provide an artificial reinforcer as a temporary, initial inducement to get the learner to perform the desired behavior. For example, a child who will not realize the natural value of brushing her teeth until much later in life might do so because she likes the taste of the toothpaste, because her mother praises her for doing so, or because her parents will yell at her for failing to do so.

While it may seem desirable to operate schools entirely on the basis of intrinsic motivation, this is actually not often practical. It is important to convince students that school in general or a specific class in particular will meet their most significant needs in the long run and to find ways to make specific tasks intrinsically motivating. However, the simple fact is that schools work most efficiently when they teach not only what interests the child, but also what experts know will be in the child's interest (Ryan, Connell, & Grolnick, 1992). Students will be much more successful in school and beyond school if they have developed internalized motivation through the processes described in Chapter 5, so that they find it rewarding to pursue worthwhile activities even when these are not intrinsically motivating. In other words, neither a child who writes a term paper "because I'll get in trouble if I don't" and one who does it "because I want to understand the subject" may initially be intrinsically motivated to do the work; but the latter is much less likely to experience negative side effects for receiving an A than the former. The undesirable side effects of artificial reinforcement are less likely to occur if the supplementary reinforcer is based on the motivational factors listed earlier in Table 5.1.

When artificial reinforcers are employed, they should always be used in conjunction with an eventual natural reinforcer. If an artificial reinforcer is used without a natural reinforcer as a back-up, the change in behavior will almost certainly be merely temporary, and there is a good chance that the artificial reinforcement will backfire by reducing the targeted behavior in the long run. In spite of the necessity of natural reinforcement, however, artificial reinforcement can often be a powerful tool to get someone to start performing a behavior which that person would otherwise not perform. Haring, et al. (1986) used artificial reinforcers to administer speech therapy to children with speech problems; but they reported that a major factor in their program was that the other children were also trained to communicate with these children - that is, to enable them to receive natural reinforcement for their improved speaking abilities.

Hennessey, Amabile, & Martinage (1989) have described an effective strategy for minimizing the possible negative effects of artificial reinforcement. They "immunized" children against artificial reinforcement by teaching children (1) to find intrinsically enjoyable aspects in whatever one has to do and to concentrate on these aspects for maximum enjoyment and (2) to cognitively distance themselves from socially imposed, artificial reinforcers and to maintain a continual focus on the intrinsically enjoyable aspects of the task. For example, through modeling and scaffolding, children would express beliefs that while it was great to get good grades, it was more important to master academic tasks because they are enjoyable and useful. Under these circumstances, the rational reinforcement actually led to increased rather than decreased motivation to engage in the artificially reinforced activities. This immunization strategy should be particularly useful in classrooms where some students need artificial reinforcers to get them started while other students are already performing the same activities for natural reinforcement.

Why Do Teachers Rely So Heavily on Artificial Reinforcement?

This chapter has tried to show that while artificial incentives may be helpful to stimulate learning among students who are initially not interested in a task, these same artificial incentives may actually have a detrimental effect on students with an initially high interest in the task.

It is simply not true that artificial, external rewards will always - or even usually - promote learning or other positive changes in behavior.

As Chapter 5 showed, an extrinsic orientation toward motivation is often detrimental to learning. (For a further discussion of the dangers of extrinsic reinforcement, see Kohn, 1993.) Students who adopt extrinsic orientations prefer easy tasks, work primarily to please the teacher and to obtain good grades, depend on others to evaluate their work, and respond negatively to evaluative feedback. In short, they perform poorly compared to learners with intrinsic motivational orientations. Therefore, to the extent that artificial rewards help students develop extrinsic motivational orientations, they should be avoided.

A major problem is that the accurate understanding of the role of artificial rewards is counterintuitive. That is, most adults firmly (but incorrectly) believe that rewards will almost always increase the productivity of those who receive these incentives. They also incorrectly tend to believe that larger rewards will almost always lead to greater productivity.

One explanation for this inaccurate perception is that adults tend to prefer children who express an extrinsic motivational orientation over their more intrinsic peers and to reinforce this unproductive pattern. This is easy to understand: extrinsically oriented children seek the approval of others (including adults), and so they are more likely than their intrinsically motivated peers to respond warmly to adult input and feedback. This process may make such children popular, but it is not good for them.

It is in the best interest of students for teachers to be wary of the seductive charms of artificial, extrinsic incentives and to promote natural reinforcement whenever possible. The following guidelines will be helpful in determining the best possible combination of natural and artificial reinforcement for a particular situation:

If it is likely to happen, learners should select their own natural reinforcers. Teachers can stimulate learners by making available opportunities for learners to pursue and receive natural reinforcement. The disadvantage of this strategy is that learners may not perceive or attain these natural reinforcers through their own initiative.
When it is necessary to artificially stimulate learner initiative, teachers should use reinforcers that are as natural as possible. In general, employing incentives based on the principles of intrinsic motivation that were listed in Table 5.1 is likely to lead to an increased motivation to engage in the reinforced activity (Lepper & Cordova, 1992).
Present reinforcers in such a way as to make them as much as possible to the logical and natural consequences of the behavior for which they are given.
Insofar as it is possible to do so, always back up any temporary, artificial reinforcer with a long-range, natural reinforcer.
Be aware that to the extent that any reinforcer is perceived as artificial and arbitrary, it is likely not only to have a merely temporary impact but actually to have the long-term impact of reducing the targeted behavior.

It is important to note that the degree to which a reinforcer is natural or artificial depends on the perception of the recipient. A reinforcer that appears natural to one person may appear arbitrary to another. It is, therefore, important to know individual learners as well as possible and to individualize reinforcement strategies whenever it is possible to do so.

Using Artificial Reinforcement Effectively...

Rolando does not want to learn to have anything to do with Cartesian coordinates. His teacher finds a computer game in which Rolando tracks down enemy submarines by entering Cartesian coordinates. The computer gives him feedback based on his entries; and by responding to the feedback and understanding how to use Cartesian coordinates, Rolando is able to destroy enemy submarines with great efficiency. The teacher has used an artificial inducement to arouse Rolando's interest in Cartesian coordinates; but this is not likely to reduce his willingness to deal with them outside the game. Rolando's reinforcement has been derived from fantasy (Table 5.1).

...and Ineffectively

Maribeth does not want to learn to have anything to do with Cartesian coordinates. Her teacher offers her five tokens for each problem she solves. Maribeth attacks the problems, solves twenty of them correctly, and trades in her hundred points for a candy bar. The teacher has used an artificial inducement to arouse Maribeth's interest in Cartesian coordinates, and this is likely to reduce her willingness to deal with them outside this token reinforcement program. Maribeth's reinforcement has not been derived from any of the intrinsically motivating factors from Table 5.1.

Praise as a Reinforcer

When I ask my students to give good examples of positive (Type I) reinforcement, they inevitably mention praise. Usually it is first on their list. However, Jere Brophy (1981a, 1981b) argues convincingly that praise is rarely used as an effective reinforcer beyond the early elementary grades. In order to function as reinforcement, praise must be (1) contingent, (2) specific, and (3) sincere, varied, and credible (O'Leary and O'Leary, 1977). Praise is likely to be effective as a reinforcer under the following circumstances:

It is linked to specific student accomplishments.
It is accompanied by a description of specific details of the student's accomplishment.
It attributes success to the student's effort and ability and suggests that similar successes are likely to occur in the future.
It suggests that the teacher has paid clear attention to the accomplishment.
It conveys to the student that the accomplishment was significant.
It is combined with a feeling in the student that he or she enjoys the activity and wants to develop skills associated with the task.
It comes as a surprise, rather than as an expected, routine part of classroom management.
It occurs in association with an occasional reprimand for non-praiseworthy behavior.

On the other hand, praise is likely to be ineffective or even counterproductive under circumstances such as the following:

It is vague and not associated with specific aspects of the student's accomplishment.
It is delivered inconsistently.
It is bland &endash; suggesting that the teacher praises almost everyone for almost anything.
It rewards mere participation, without considering the quality of the student's performance.
It gives the student little information about the relevance of the accomplishment to outcomes that he or she considers worthwhile.
It attributes success to ability alone or to external factors such as luck or level of difficulty.
It encourages the learner to shift motivation from internal to external factors &endash; from doing the behavior for its intrinsic merits to doing it to please the teacher.
It is delivered in such a way as to appear ironical or to embarrass the child in some other way.

Brophy (1981b, pp. 16-18) identifies several other functions that praise serves other than reinforcement. (Its most frequent function is to signify that an activity is finished and it's time to move on to something else.)

Brophy's research indicates that praise does not occur very frequently in the classroom. This is fine with him, because a higher incidence would not be very useful anyway. It would simply be impossible for the teacher to distribute the massive amounts of praise that would be necessary to use it as a primary means of classroom management. As Brophy points out, "A teacher can concentrate on a few specific behaviors for the class as a whole, or on a large number of specific behaviors for a few individuals. Beyond this, however, the teacher must function by obtaining the general cooperation of the students rather than by continually reinforcing their specific behaviors." (page 20) This does not negate the fact that reinforcement for specific behaviors is important. Rather, these specific behaviors must be reinforced without requiring constant attention from the teacher. This works best when the learners reinforce themselves.

To the extent that praise is artificial, it possesses the same disadvantages as other extrinsic reinforcers, such as money and candy. In some cases, praise is neither pleasant not unpleasant. At its worst, praise may function as a form of punishment by making a child conclude, "I've done so badly that the teacher has to praise me for doing a worthless job." It is the attribution that the recipient assigns to praise that determines whether or not it will be reinforcing.

All this is not to say that praise is unimportant. Praise can be an effective reinforcer, but only if it is employed properly. However, there is no evidence that encouraging teachers to praise more often will enhance learning (Black, 1992). On the other hand, encouraging teachers to have better rapport with their students is important. It is the improved rapport and classroom management, not necessarily the praise, that improves instruction in a general sense. If praise is to be used as an effective form of reinforcement, it should be given relatively rarely and for concrete activities that actually deserve praise. Ideally, praise should focus the attention of the learners on their own task-relevant behavior, enabling them to praise themselves when they perform correctly.

Where do you want to go next?

Introduction to Reinforcement
Artificial and Natural Reinforcement <<You are here>>
Shaping New Behaviors
Applied Behavior Analysis
Token Reinforcement
Skill Training and Self-Management
Comparison of Types of Reinforcement
Eliminating Behavior through Reinforcement
Details of Reinforcement
Summary
Answers to Quizzes
Annotated Bibliography