THE BASIC ELEMENTS OF SHAPING: THE STRATEGY FOR DEVELOPING AND MODIFYING SPORTING BEHAVIORS

[Extracted from Rushall, B. S., & Siedentop, D. (1972). The development and control of behavior in sport and physical education. Philadelphia, PA: Lea & Febiger. (pp. 137-147).

This reprint is of the first known exposition of operant shaping as a teaching strategy for coaching and movement instruction. It is as relevant today as it was almost 30 years ago.]

The previous chapters have discussed the general features of operant conditioning as they apply to behavior control in sports and physical education. These features can be drawn together to produce a strategy for teaching new behaviors and modifying established behaviors. The strategy is referred to as shaping. Physical education research has principally addressed itself to limited aspects of this topic with the consideration of such items as part versus whole learning, speed versus accuracy, etc., but seldom has a coherent strategy for producing change been outlined. Lawther has come closest to outlining a method with his "gross-framework" model for teaching physical activities to beginners (Lawther, 1968, page 65).

It would be desirable to control all the factors which maximize the learning of physical skills. This is rarely possible in teaching and coaching. To form a pragmatic method of teaching, it is necessary to select the features of control that are accessible, relevant to the environment, and feasible for use. The adaptations and efficiency adjustments of a performer occur at some covert level and quite often at a subconscious level. Conscious performance strategies of learners are particularly individualistic. Teachers and coaches are usually unable to control such strategies although they can be influenced with methods of prompting and guidance (verbal instruction, demonstrations, film). The potential for changing and developing the topographies of behavior lies generally in the manipulation of the external environment, particularly the stimulus setting and consequences.

It must also be stressed that there is a difference between teaching physical skills and controlling behaviors in the same environment. Teaching addresses itself to developing the topography or rate of occurrence of a new behavior or changing an already established pattern. Controlling concerns itself with the manipulation of repertoires of established behaviors that are required in an environment. The shaping strategy is concerned with the development of specific behaviors.

The shaping strategy involves:

  1. the definition of a terminal topography,
  2. the sequencing of steps of closer approximation to the terminal behavior,
  3. the use of primes and prompts to produce performance variations, and
  4. the use of reinforcement.

Shaping consists of reinforcing closer and closer approximations of a desired terminal behavior. It is unlikely that a correct response will be emitted on a first attempt so a behavior short of the final behavior form must be reinforced. With subsequent emissions of the behavior, the requirements for reinforcement are made more strict and reinforcement is only provided as the behavior more closely approximates the desired act. The criterion for reinforcement changes from being seemingly lax to being stringent for the final performance. In shaping, as in all reinforcing contingencies, it is essential to reinforce behaviors before other responses intervene and disrupt the pattern. During shaping, reinforcing consequences not only strengthen the particular responses but also increase the probability that a closer approximation of the final behavior will occur. This is the main reason that shaping works. As new approximations are reached and reinforced, aspects of earlier behaviors are extinguished. Skillful shaping consists of selecting the right responses to reinforce and in knowing how long to reinforce each approximation before moving to the next sequenced step. It also requires clear definition of the terminal behavior and the planning of the sequential steps so that transitions are easily made. The administration of reinforcers is the main difficulty with shaping. It is for this reason that it was suggested elsewhere (Rushall & Siedentop, 1972, pp. 106-107) that instruction be undertaken in small groups where reinforcers can be provided continuously and contingent upon behavior. The first instance of the attempted behavior is far removed from the final desired product but it is reinforced. With instruction and reinforcement the behavior successively improves by more closely approximating the final topography.

IMPORTANT CONSIDERATIONS FOR SHAPING PROCEDURES

The Definition of the Terminal Behavior

Before any developmental procedures can be implemented, it is necessary to define the terminal behavior that is to be shaped. Such a definition would include all the behavioral elements which need to be recognized and understood. Each element must be described in observable and measurable terms.

General behaviors such as rule-following and competitive effort are usually defined in terms of their general features and effects. General features describe the method of how one follows a rule or how one asserts himself in competition. These features do not require a defined topography but rather a compliance with a set of procedures. For example, the definition of an attending behavior might be:

Attending behavior. The pupil should arrive on the gym floor before the time scheduled for the commencement of class. He should be appropriately dressed for the activity that is to be instructed.

In this definition two observable criteria are listed, arriving at a specific time and the manner of dress. If these characteristics are not displayed then the pupil will not be assessed as attending the class. A behavior which also requires an effect should include that effect in its definition, e.g. an assertive behavior may stipulate obtaining a heart rate of 160 beats per minute. The criterion for an adequate definition is that it should allow any observer to recognize a response of the operant class under consideration. The explicit definition of topographical aspects is not usually necessary for general behaviors as the procedures and effects of the behavior are the features of importance. This is what differentiates a general behavior from a skilled behavior.

On the other hand, in order to define a skilled behavior it is usually necessary to consider topographical features, procedural features, and effects. For example, the skill of tackling requires the head, arms and shoulders to be placed in a particular relationship to the object to be tackled (topographical features), it requires the tackler to undergo certain preparatory and execution procedures (procedural features), and it requires that the opponent be put on the ground (effect). These features differentiate a "tackle" from someone inadvertently tripping over a defensive player (the procedures and topography are absent), from missing a "tackle" (the effect is absent), etc. These criteria define a "tackle." They allow an observer to know what a "tackle" is, and they provide a means by which a performer can be credited with having performed adequately. It is usually necessary to consider three aspects of the skill in formulating a definition:

  1. Temporal factors. Timing the onset of the response, coordination of muscular activity and the anticipation of time-dependent events are examples of temporal factors which occur in some skills.
  2. Spatial factors. The relationship of the performer to various features of his environment such as a target, a ball or an opponent are examples of some spatial factors which occur in some skills.
  3. Proficiency factors. Proficiency can be further broken down into three components, (1) the requirements for accuracy (as in pitching or kicking), (2) the requirements for productivity (as in height jumped, distance thrown), and (3) the requirements for efficiency (the optimal expenditure of energy resources over time as in an endurance run.

Most skills have components Of all these factors in their structure. These features should be considered in the initial development of the topographical elements of the definition.

In Chapter 6 the concept of chaining was explained. If the chain is explained in macro terms, for example, a volleyball spike consists of anticipating, preparing and spiking behaviors, then each of these members of the chain are discrete behaviors. To perform a volleyball spike at a high level of proficiency, each member of the chain would have to be executed well. It is possible to shape each of them. In this regard one does not shape a chain of behaviors when initially instructing a skill. The Procedure is to shape each of the macro units, and then shape the chaining of these units to finally produce the complex behavior. It is important to recognize the fact that the behavior that is to be shaped should be small enough in terms of acquisition time to indicate progress of achievement of the final complex behavior. In this regard it would be better to teach anticipation (the utilization and recognition of certain stimulus situations), preparation, and the spike (gathering, jumping, extending, hitting, follow-through) as separate units rather than shaping the whole skill from the outset. This procedure allows for the use of progress charts, etc., as reinforcers as it does set defined developmental stages in acquiring the terminal complex behavior.

Skilled motor behaviors usually fall into one of four instructional classes:

  1. Some motor behaviors are highly segmented and best taught as discrete units. For example, typing in the beginner stage consists of a large repertoire of small skills. Each letter typed is an individual response which needs to be developed through meaningful practice. With practice and reinforcement, the nature and size of the performance unit change from letters to words and then to phrases. In the final stages the typing skill is smooth and flowing which is of markedly different appearance to the requirements of the beginning skill.
  2. Some motor behaviors are performed only as whole units. Woodworth (1958) and Fitts (1964) indicated that the basic behavioral unit for meaningful practice is the two-phase motor unit (a preparatory act and then the act itself, e.g., a crouch and then the jump). Several sports skills, e.g., pitching, hitting a golf ball, a somersault, are best taught as an attempt at the skill itself. Such skills are usually very short in duration and the resulting performance information is used to adjust the next trial of the skill.
  3. Some motor behaviors are practiced as whole units but are progressively changed through instruction. These behaviors are highly repetitive tasks. Instruction serves to alter the topography of the skill while it is in progress, e.g., changing a running action, learning to row.
  4. Some motor behaviors are briefly introduced in small units and then practiced as a whole unit. For example, a beginning swimmer is shown how to kick and move the arms. These units are practiced before entering the water. In a short while these segments are combined and practiced as a unified activity from then on.

The implication from this brief discussion is simply that one needs to know the behavior that is to be shaped. It is possible to shape a behavioral element or a large-scale activity. The procedural steps for shaping are the same regardless of the magnitude of the behavior. In the final stages of any behavior development the behavior must be practiced in its entirety. When a segment is practiced and reinforced it should be qualitatively the same as in the total behavior. It is the teacher's, or coach's decision as to what is the unified behavior that is to be developed.

Priming Behaviors (Guidance)

When a new behavior is to be shaped, it is necessary to start the procedure with a behavior which is at least allied to the terminal behavior. The procedure provoking such a behavior is called priming, prompting, or guidance. Primes act as discriminative stimuli for the performer as they indicate the appropriate behavior to be emitted. Holding (1965) presents a clear and comprehensive review of the research associated with this topic.

Methods of physical, visual, and verbal guidance are all directed towards limiting the occurrence of errors in the stages of acquiring a behavior. Errors appear to be learned because people who commit them tend to repeat them. Erroneous responses have to be suppressed in a later stage of the behavior development process. This problem can be avoided if ways can be devised to prevent people learning or experiencing erroneous responses, for example, through the use of programmed instruction or training devices.

There are usually four types of guidance procedure. Physical restriction prevents the occurrence of errors as the performer is blocked from making an overt incorrect response. Floats for beginning swimmers and a harness for practicing gymnastic activities are two examples of physically restricting devices. Forced-responding dictates to the performer how one should perform. Physical manipulation of the limbs in showing how to do a crawl stroke kick is an example of forcing a response. The subject does not actively initiate the response. It is possible that forced-responding could produce some intrinsic kinesthetic cues. Another method is visual guidance. Methods of completing a response or task are presented in some visual form, for example, plans, maps, a series of photos. The final method is verbal guidance. This can enhance performance by communicating information other than pure guidance.

Research results have produced a series of findings which can be of use to the teacher and coach when teaching behavior chains (i.e., serial tasks such as dance steps, gymnastic routines, and football plays). Early guidance in the instructional process helps preserve response flexibility in the face of frustrating or anxiety-provoking conditions. Some teaching situations are quite stressful for the performer, for example, the adult learning to swim usually suffers some social embarrassment and often displays "fear" reactions. It is important for the teacher to give adequate guidance in such circumstances. Individual instruction in each developmental step is usually required. It has also been found that if there is a basic change in the instructional procedure after training has been initiated, by either beginning or ending guidance, learning is retarded. This is yet another reason for proposing shaping as the instructional strategy.

Physical Restriction. The possibility of using devices for instructional purposes has been discussed elsewhere (Chapter 3 in the section on performance information). Such devices limit the potential for making errors or experiencing unpleasant circumstances. Usually these devices are provided as a safety precaution (e.g., swim floats prevent sinking and the safety harness in gymnastics reduces the prospect of injury). The consequence of using these devices which facilitate instruction is of secondary importance to their traditionally implied safety role. The scope for researching instructional devices for physical skills is quite broad.

Forced-responding. Cooperative manipulation of an individual provides kinesthetic experiences which are not offered through passive methods of guidance (visual or verbal). Manipulation refers to the external control of a performer's anatomy such that the locus of action approximates a terminal behavior, for example, moving the legs in a flutter-kick action in teaching swimming. The individual who is undergoing manipulation must be cooperative. If an individual is not motivated to perform the behavior, manipulation has aversive characteristics and learning is retarded. Manipulation is particularly useful in teaching retarded and sensorily deficient individuals. It is one of the least researched methods of priming.

Visual Guidance. There are a variety of methods which are suitable for priming behaviors. Demonstrations are perhaps the most common form of visual guidance. They provide standards for a performer which are essential for activities which provide little intrinsic information feedback. They provide a basis for imitation which serves to narrow down the scope of possible trial-and-error responses. Demonstration should be supplemented with meaningful reinforcement for appropriate responses. It is a major method for priming behaviors when verbal instruction fails or where experiences or some senses are lacking (e.g., raw beginners, the deaf). In demonstrating a behavior, one should attempt to present it as the learner sees it. When verbal guidance is used in conjunction with demonstration, content should be aimed at the features which are important to the observer. This will enhance the number of relevant cues in the demonstration.

Visual aids are popularly accepted methods of guidance and priming. This is mainly because of the concerted efforts of the commercial enterprises which produce them. Pictures, graphs, etc., may serve as variety features in a lesson or book but what evidence there is does not give much support to the value of their use. A film produces information at a fixed rate and evidence suggests that films fall short of their promise. The best procedure is to supplement films with question and answer procedures rather than allow just passive viewing. However, displays which provide important information do enhance learning and the initiation of a desired response. They should highlight essential important facts, for example, a picture of a kicker with arrows pointing to the important features is much more effective than the picture alone. If a display is used to guide a task (as in learning to type, run a machine) it is best to have the display spatially oriented with the task. For example, to teach a series of dance steps it would be better to have the routine marked on the floor rather than illustrated on a board. The display would be oriented directly to the activity. Similarly, it would seem to be better to learn various plays for football by marking the movement patterns on the ground rather than explaining them on a board as is the usual procedure. The ground markings restrict the possibility of making errors and also produce a precision factor into the plays at a much earlier stage than is usual. Where tasks have certain perceptual difficulties (something happens too quickly for the naked eye to perceive or it is too large to comprehend) displays can be utilized. The tasks can be magnified either in size or speed (slow motion movies) or they can be "minified" (a football field is scanned by a TV camera and seen as a whole).

Verbal Guidance. Verbal instruction is the more traditional form of priming responses. It is limited in some respects to the ability of the listener to interpret instructions and to the speaker's verbal ability. Verbal instruction is actually an intervening process between the observation or conception of the terminal behavior by the instructor and the understanding of the student. Verbal instruction by itself serves as a translation process and consequently loses information as it is implemented. It requires more preparation for its use than is traditionally afforded it in instruction. It embraces a variety of purposes which include:

  1. Directions on how to perform a skill.
  2. Directions on how not to perform a skill.
  3. Reasons for why a skill should be performed in a certain manner.
  4. Verbalizations which focus attention on aspects of a skill for feedback.
  5. Verbalizations for motivation (after Lawther, 1968).

Verbal instruction becomes more important and useful as the performer becomes older. It requires the teacher or coach to develop a mastery of the skill vocabulary particularly when highly skilled performers are involved.

The effect of instructions depends upon the form of the instructions and the nature of the task. Instructions which over-elaborate or inundate the listener with information can retard behavior development. For example, in highly skilled athletes a preoccupation with analysis and cognitive control may retard acquisition and detract from performance. The role of the prime is to act as a discriminative stimulus for an appropriate behavior. Confusion and ambiguity in the stimulus will decrease the probability of the appropriate response being emitted. Attempts to impress others by using big words and complex concepts tend to retard learning.

Verbal pre-training is also a form of verbal guidance. Saying what will be done can be used as a preparation for behavior. The recital of the steps for performance in the initial stages of instruction aid acquisition provided the words help to make perceptual distinctions and that the verbal response does not interfere with the behavior. For example, in sailing the procedure for a skipper to follow when wishing to "go about" is clearly defined and is usually learned through on-shore training. When the actual response is to be made the novice skipper rehearses the verbal checklist prior to issuing commands. A typical rehearsal consists of:

  1. warn the crew,
  2. give the ready signal,
  3. issue the command "go about,"
  4. hard tiller,
  5. change position, (
  6. pull in the main sheet, and
  7. steady the tiller.

This series of steps guides the beginner to execute a complex skill and decreases the possibilities of errors in performance. Verbal pre-training is suitable for the early stages of learning in tasks which:

  1. differentiate a set of motor behaviors (e.g., the steps concerned with changing gears in a manual car),
  2. which differentiate between a number of stimuli and the appropriate action (e.g., if X then Y, but if Z then X), and
  3. where the verbal response means the motor action.

The content of verbal pre-training commands must have action counterparts and must direct the sequencing of the activity. In essence, verbal pre-training consists of learning a chain of verbal events which later act as a series of discriminative stimuli for a motor chain.

In general, words can be used to supplement behaviors in the initial stages of behavior development. As the acquisition process proceeds, the role of verbal priming and guidance should be faded out (stimulus fading).

The above priming techniques can be combined to provide more effective methods of information transfer to prompt a particular behavior. Combined forms of sensory input produce more information transfer than singular methods. It would be best to try to combine as many forms of sensory input as possible to produce a behavior that is closer to the terminal behavior than is normally provided by single forms of priming. Audiovisual instruction (verbal instruction plus demonstration) is a form of combining priming methods. Other possibilities exist for combination but have not been extensively researched.

Step Sequences

The difference between each step in the shaping process is important. It is relatively rare that an individual can display a terminal behavior after priming without going through some series of approximations. One should not expect to reinforce only the correct skill in teaching physical activities. Miniature segments of a complicated technique need to be reinforced to firmly establish behavior patterns. A reinforcement after a long period of practice which includes performance variations generally does not influence the details of a skill.

The steps in the shaping process should be sequenced so that each step progresses with equal ease to the next. The terminal topography may look to be extremely complex and difficult, but when it is finally reached it should be executed as easily as the first step. If there is difficulty in transition from one step to the next, then faults in the step sequencing (programming) are most probably present. Stimulus ambiguity, confusion, inadequate priming, instruction, or direction, or aversive contaminants may cause disruption. Each step in the programmed set must be overtly attempted and superimposed upon the previously reinforced step. Program steps need to be fairly small and must be organized into an effective sequence. The order of complexity, difficulty and logical progression are factors often considered in developing teaching and coaching schedules. The step size should be small enough to allow immediate reinforcement. This raises the frequency of reinforcement to a maximum while the possibility of the aversive consequences of committing errors is minimized.

The amount of change in behavior that is demanded of the performer in each step must be weighed against the need to maintain the behavior at a given strength. If the step transition is too difficult and no reinforcement is forthcoming, then already reinforced behaviors will be extinguished and no new behavior will be reinforced. If the step is not difficult, a new behavior will be strengthened while components of the old behavior will be extinguished. It is important that the coach and teacher select the correct hierarchy of approximated responses for reinforcement, and then know how long to reinforce each approximation. It is not advisable to reinforce an inadequate behavior when trying for a better response. Such a reinforced behavior may be competitive to the desired new behavior. The hierarchy of steps that is planned by the coach and teacher must be graded, and reinforcement should only be given when each step is correctly executed. When an athlete does not appear to be able to perform the next programmed step in the sequence, it is most probable that the step size has been too large. The coach or teacher will then need to redesign the step progression so that each unit can be performed by the individual. In the instructional procedure, the teacher or coach should return to the previously successful step and then proceed with the new sequence.

It is likely that each step will have to be primed in some way. The reason that priming each step is desirable is that immediate cueing heads off incorrect anticipatory responses (primes are discriminative stimuli). Progressive prompting is a very effective teaching method when it is combined with contingent reinforcement.

Determining when to progress from one step to the next is a decision that has to be included in step-formulation. If one correct performance occurred it could be because of a chance happening or actual learning. To avoid the former, which would likely result in errors in the next step, it is usual to develop a progression criterion that requires more than one correct response to be performed consecutively. When a step's performance is errorless for several trials, the reliability of the learning at that step is demonstrated and a chance occurrence can be ruled out. In practical terms, a common progression criterion is a minimum of three consecutive correct trials in a step being performed. If two correct responses occurred and then an error was made, the requirement for three in-a-row would need to be restarted. It is important to be rigorous in the application of step-progression criteria.

Schedules of Reinforcement with Shaping

Schedules of reinforcement must be prepared for two phases of the shaping process, the step and the terminal behavior. Each step must be performed a number of times to eliminate any chance occurrences. Each step must be reinforced sufficiently to have enough strength to be performed on a few occasions without reinforcement when step changes are introduced. The most appropriate schedule for steps is a burst of continuous reinforcement (CRe) which is then stretched into a low variable ratio (VR) schedule. All that is required of the performer is to be able to perform the previous step while attempting to add or replace the new behavioral elements of the next behavior. If a skill is considered as consisting of a set of behavioral elements and the shaping process is an agradation of those elements, then the already reinforced elements which are required for the terminal behavior need to be consistently exhibited with each subsequent step. A reinforcement schedule is required to partially establish the behavior of each step.

When the terminal behavior is exhibited, the shaping process is completed. What is then required is for the terminal behavior to be firmly established and maintained at a desirable rate of occurrence. The coach and teacher must provide a schedule of reinforcement which will fix the behavior as a permanent feature of the individual's behavior repertoire. CRe changing to variable ratio/variable interval (VR/VI) reinforcement is the appropriate schedule here. An added feature that is commonly required for the substantiation of terminal behaviors is "booster" or revision sessions where further CRe-VR/VI schedules are administered although in successively diminishing amounts. It is also often necessary to repeat the last few steps of a skill development program for several sessions. This procedure helps to diminish variations in performance which occur in the early stages of high level skill acquisition.

The final decision for effective development and control is to develop the amount of stimulus control that is needed. Certain skills need to have highly developed discriminative control. Skills which require decisions as to how to react fall into this category. Most sports activities are in this class. The procedures for developing stimulus control have been discussed elsewhere in this text.

Reinforcers

Reinforcers appropriate for operant conditioning have already been discussed. The most useful reinforcer for the shaping of skilled behaviors is knowledge of results. However, it is often a wise procedure to have several other reinforcers on hand as back-up consequences in case of possible satiation in long shaping procedures. Knowledge of results is particularly powerful when it is tied to a contingency management plan (see Chapter 9). Younger persons are more susceptible to a greater variety of reinforcers than are older individuals.

Shaping procedures are used for developing behaviors where the terminal behavior does not exist. This embraces two areas:

  1. the teaching of a new behavior, and
  2. the changing or superimposing of aspects of a behavior. The models for each of these tasks are similar in that they do follow the operant, shaping-by-approximation concept. They differ in some aspects of the strategies involved. It is because of these differences that the two are presented separately.

References

  1. Fitts, P. M. (1964). Perceptual-motor learning. In A. W. Melton (Ed.), Categories of human learning. New York, NY: Academic Press.
  2. Holding, D. H. (1965). Principles of training. Oxford, England: Pergamon Press.
  3. Lawther, J. (1968). The learning of physical skills. Englewood Cliffs, NJ: Prentice-Hall.
  4. Rushall, B. S., & Siedentop, D. (1972). The development and control of behavior in sport and physical education. Philadelphia, PA: Lea & Febiger.
  5. Woodworth, R. S. (1958). Dynamics of behavior. New York, NY: Holt, Rinehart, and Winston.

Return to Table of Contents for this issue.