[An abridged version of Sherman, C. A., & Rushall, B. S. (1993). Improving swimming stroke using reverse teaching: a case study. In W. K. Simpson, A. D. LeUnes, & J. S. Picou, (Eds.), Applied research in coaching and athletics annual 1993. Boston, MA: American Press.]


Teaching with reverse progressions is one behavioral coaching procedure described in the sports-related literature (Chelladurai & Stothart, 1978; Rushall & Ford, 1982) and sometimes called backward chaining or backward shaping. With reverse teaching progressions, the behavior to be trained is defined by the trainer as comprising two or more components, which are serially ordered. Training begins with establishment of the final or terminal component. After a criterion performance of this position is reached, training proceeds to a sequence that comprises the immediately preceding component, followed by the terminal component. Once this step or link has met criterion performance the next preceding component is added and so on, until the learner performs all the components in smooth succession always finishing with the terminal component.

Most discussions of reverse teaching progression in the coaching literature have consisted of descriptions of the progressions but not of data collected during an application of the procedure. For example, Chelladurai and Stothart (1978) suggested backward chaining as an acceptable procedure for basketball skills (lay-ups) and volleyball spiking technique. Dusault (1986) also suggested that volleyball spiking technique could be enhanced in the learning phases with backward shaping procedures. However, no empirical data were presented in these papers. To that end, the purpose of the present study was to plan and implement a reverse progression teaching strategy for the repetitive skill of front crawl swimming. This application was intended to describe in detail the procedures, treatments, and results involved in modifying an already successful swimmer's hand and arm technique.



S was an elite Canadian 16-year old female swimmer who had observable topographical deficiencies in front crawl swimming technique. She had 10 years of competitive swimming experience, and had qualified for finals at the national championships as well as being ranked as the best in a number of events for her age group. She was training seriously at the time of this study.

Setting and Sessions

An indoor heated 50-meter pool was used. Individual coaching and strategy implementation were performed by the senior author. After an initial observation session, three more sessions were conducted during regular training times in a sectioned area of the pool. Each session took approximately 40 minutes to complete.


Swimming goggles were worn by both S and investigator. A pull-buoy to restrict the leg movements and a kick board to restrict unnecessary movement of the non-moving hand were used during the early stages of instruction.


Behavior of interest. The front crawl arm stroke technique was chosen as the behavior of interest in this study. Initial assessment of S's swimming technique was paired with the corresponding correct arm and hand components described by Maglischo (1982). This resulted in a list of incorrect components in the swimming movement pattern of S's arm and hand. Definitions of the correct technical elements and the observed errors were formed.

Preliminary session. A single preliminary session was used to make initial observations of S. S's coach served as an independent observer. As S swam warm-up and race pace 25 meter lengths, observations were made from the pool decking both laterally and frontally. Further observations were made from beneath the water surface. As front crawl arm stroke technique faults were detected and agreed on by both observers they were recorded.

Both warm-up and race-pace swimming were observed to assess stroke rate and speed. S performed 10 repetitions of 25 meters of swimming first at warm-up pace and then at race pace (20 repetitions total). The time, using hand-held electronic stop-watches, and number of strokes taken for each length were measured and recorded by both observers. This evaluation was repeated after the completion of the reverse teaching progression procedure so that a comparison of pre- and post-instruction data was possible.

Reverse teaching progression sessions

Three 40 minute sessions were used to apply the reverse teaching strategy. In all three sessions the investigator and S were in chest deep water in a sectioned-off lane of the pool.

First session. S was given a preliminary explanation regarding the aims and goals to be achieved. No mention of her incorrect technique was made. However, it was noted that possible changes to her technique could improve her swimming efficiency. S was then prompted with demonstration and description of the first teaching step. This was the correct terminal component (TC) of one arm and hand which was to be engaged in by S at the end of every step of the instruction (see Table 1). S was then required to hold this arm position while lying horizontally on the water surface. In this position, and most other developmental steps throughout the procedure, S was required to use a pull-buoy to restrict leg movement and hold a kick-board to restrict unnecessary movement of the free hand. The localization of only one moving arm allowed the swimmer to concentrate on and experience more acutely the sensations that accompanied correct arm and hand movements.

When the investigator was satisfied with S's demonstration and familiarity of the correct finish position for each hand, she was prompted to perform component A (Table 1) followed immediately by the terminal component and to continue this sequence for 25 meters of the pool. The linking of component A and the terminal component (A+TC) in that order, was the requirement for step 1 (Table 2) of the reverse teaching progression. Step 2 (Table 2) consisted of the same as step 1, however, with focus only on the left hand (B+TC). Step 3 (Table 2) reverted back to the right hand (C+A+TC) and so on until all 13 steps were completed.


Behavior Content
Arm has completed backsweep and hand touches thigh. Little finger is closest to water surface.
Half insweep and full backsweep with right hand.
As for "A" with left hand.
Half downsweep with right hand.
As for "C" with left hand.
Full downsweep with right hand.
As for "E" with left hand.
Right hand entry.
Left hand entry.
Half arm recovery with right hand.
As for "I" with left hand.
Release and recovery with right hand.
As for "K" with left hand.
Full arm stroke using both arms.

Note: Each step included the previously completed components and always finished with the terminal component (TC).


Component Order
continual swimming using all components

For the partial arm cycles required in most of the early steps, the action was stopped at the finish position (TC) so that the sensation of the body gliding through the water as a result of the action could be experienced. Recovery of the moving hand after reaching the finish position was performed under water, hence a slight delay occurred after each body glide. During the demonstration of the initial 25 meters and all other swims, S was viewed from beneath the water surface so as to observe any errors. There were no errors detected during the first trial and on completion S was given social praise and performance evaluation that constituted positive knowledge of results. S's gradual improvement in velocity maintenance and glide distance also served as a more immediate consequence of changes in stroke topography.

After S had experienced the initial requirements she was told that if she could perform four consecutive errorless 25 meter lengths for step 1 (Table 2) A + TC, then the next step would be introduced. She then attempted this requirement and after completing the first length an error was detected and corrected by the investigator. A verbal prompt for the required hand position was given in order to correct the fault and the count of trial lengths was restarted. The investigator recorded both trial errors and successes using pencil and pad kept near the side of the pool. The first session continued with S progressing through the first four steps, always requiring four errorless 25 meter lengths in succession before moving on to the next step. Therefore, the correct execution of each new instructional component, as well as the ability to correctly demonstrate all the previous components of the technique, were necessary prior to the addition of the next step. This meant that the criterion for a successful trial became more complex and demanding as the progression advanced. Only one other error was detected in this first session which occurred on the second trial length of step 2 (B+TP).

Second session. This occurred at the same time the next day. Once again S was required to fulfill the same criteria as the previous session. Steps 5 to 8 (Table 2) with components E to H (Table 1) were included and S made four errors.

Final session. This also occurred at the same time the next day. Steps 9 to 13 (Table 2), including components I to M (Table 1), were completed and only three errors were recorded. Step 13 (Table 2) with component M (Table 1) employed the first use of both arms moving in the normal swimming fashion for the 25 meter lengths. Kicking was not allowed at any time. This only occurred outside the teaching sessions when the swimmer resumed normal training. At no time were restrictions made on breathing. However, it was preferred if S could cover as much distance as possible for each length on the minimum number of breaths. This was to ensure that S was watching her hand movements as closely as possible.

Follow-up session. The day after the instructional procedure was completed, the initial testing for time and stroke count over 25 meters at warm-up and race paces was repeated. The co-observer evaluated the two performance features with the intent of establishing the reliability of the primary investigator's observations. Observations were again made regarding S's arm stroke technique for front crawl and all data were recorded.


Inter-observer Analysis

The initial evaluation session resulted in producing a status description and list of S's inefficient technique elements for front crawl swimming. Those observations were verified by the coach. After Session 3, the observers agreed and concluded that none of the original erroneous elements were evident in the swimmer's stroke. Those observations served as independent verifications of the investigator's objectivity and accuracy. The observers independently measured the time and number of strokes for each 25 meter swim in the post-treatment evaluation swims. The results recorded by the coach were in perfect agreement with the measurements taken by the first author.

The number of strokes decreased from 19 to 16 (15.8%) at warm-up speed, and 22 to 20 (-9.1%) at race speed. The time for a 25-meter length decreased from 20.1 to 19.3 (-4%) seconds at warm-up speed, and 16.4 to 15.9 (-3.1 %) seconds at race speed.

Consistent and successful learning was evidenced. Few errors occurred (two in Session 1, five in Session 2, and three in Session 3) and were usually followed by immediate correction. Only in step 13 with component M (hand release) was an error repeated (twice). Overall, 10 incorrect trials among the total 75 trials were recorded.

Social Validation

During the implementation of the instructional sequence, S admitted to feeling a change in the stroke technique. She reported increased amounts of local arm fatigue and after-training soreness, two indications that different muscles or parts of muscles were being used. After several training sessions, the discomfort disappeared. The coach remarked favorably about the unobtrusiveness and speed of the technique alteration. He stated that he would attempt to employ the strategy with other athletes and would recommend it to other coaches. The procedure was deemed to be acceptable to the athlete and the coach and to have social validity.


This investigation aimed to demonstrate the effects of teaching behavioral components in a reverse progression. The common and popular forward instructional progressions diminish in effectiveness when they are long or involved or are used to alter an existing technique. Reverse progressions are supposed to be more effective for changing behavior as well as facilitating the occurrence of very few errors.

Traditional swim coaching attempts to instruct techniques in a forward progression. The hand entry is usually taught first, then followed by the attempt to exert productive forces underwater, followed by aspects of the recovery. Progressions are of varied numbers of steps and step detail. It is easy to understand why coaches adhere to forward progressions since film sequences of good stroking follow that sequence and it seems a "logical" order. However, because self-evident justifications exist for adopting a progression that does not mean they are the best for ordering the presentation of instructions and steps.

The swimmer in this study altered a considerable number of front crawl technique features that had existed in her stroke for a long time. Those changes occurred in a very short time period. When one reflects on the number of items and their anticipated difficulty for changing, what was demonstrated is quite remarkable. The topography or technique of the underwater stroking action was altered in three 40-minute instructional sessions.

Essentially, the instructional strategy was a "package" of teaching components. Some of the most note worthy aspects of that package are described below.

  1. There was only one swimmer in the instructional group. This allowed for an intense instructional session that facilitated the athlete's total concentration. This period of dedicated application may be a very formidable aspect for producing dramatic learning in a swimmer.
  2. Each trial produced a rich set of consequences that directly reflected the intention for learning on that trial. The power of contingent and continuous reinforcement in the form of praise and knowledge of results cannot be overestimated. The situation that confronted S was markedly different to that which usually exists in instructional periods in a group training setting. In the more usual circumstance reinforcement is rare and not always contingent upon behavior. The variation and inconsistency in its presentation produces no consistent theme of development leaving most participants to react to natural contingencies which exist in the swimming environment which, in turn, produce the adoption of convenient movement patterns. Often those movement patterns are not the most efficient actions. Isolated attempts by coaches to instruct technique changes usually have a transitory effect for they do not establish consistent sequences of reinforcement that are contingent upon repetitious forms of response.
  3. The reverse teaching progression always presented new components to be learned as the first element in any practice trial. This allowed the swimmer to focus completely on the new instructional challenge and then follow it by doing what had been completed successfully in the previous steps to finish at a particular stroking posture that could serve as an intrinsic reinforcer indicating completion of a successful performance. This is one of the major strengths of reverse progressions, the swimmer always knows what has to be done to be successful after a new component is attempted. That does not occur with a forward progression teaching.
  4. The progression was implemented according to the usual criteria for successful effects. The steps were of sufficiently small size to allow S to easily progress without making repeated errors. Seven of the ten errors occurred on the first trial of a step. This may have been due to the inadequacy of the prompts that were administered. The experience of being overwhelmingly successful and correct in learning attempts is a critical feature of effective instruction. The number of correct trials is associated with performance improvement more than any other topographical feature of skill instruction (Ashy, Landin, & Lee, 1988). The manner in which this reverse teaching progression was implemented facilitated a very large number of correct trials and very few errors.
  5. The alterations achieved in the swimmer's technique were components that occurred underwater. Since that is where propulsion is generated they should have the greatest effect on performance. To produce these changes it was necessary that the swimmer be analyzed under water. Few, if any coaches instruct from this viewpoint. They tend to teach from the deck and emphasize those features that are readily observable (e.g., the recovery, entry, and early pull). Thus, the emphasis on very influential performance factors that occurred in this investigation could be a significant reason for the dramatic performance changes observed.

These features are some of the basic ingredients of effective teaching that were demonstrated in this investigation. However, one important feature was not entertained, the stretching of the reinforcement schedule to produce response permanence once the total terminal behavior was exhibited. S's lack of availability prohibited this from occurring. If any reader were to attempt to implement reverse progression teaching strategies, it is necessary that he/she add this vital feature.

The objective improvements in S's swimming stroke were undeniable. The strategy accurately altered an existing technique to include new and more correct elements. Observations of pre- and post-intervention performance characteristics verified this result. Not only did the skill topography change, but so did performance. At both warm-up and race speeds the distance covered per stroke was improved as was the time for the 25-meter distance. These effects were verified by independent observation as well as the subjective comments of the swimmer and her coach.

If only the time for the test distance was improved, it would be reasonable to assert that the change was due to any number of causes. However, the speed change was yoked to a reduction in strokes taken over the set distance. That means the efficiency of swimming was improved at both warm-up (sub-maximum) and race (maximum) speeds. At sub-maximum speeds, the stroking efficiency improved by 15.8 percent which resulted in a 4 percent improvement in speed. At maximum speeds, stroking efficiency improved by 9.1 percent and speed by 3.1 percent. These effects are very noteworthy in an elite swimmer. They meant that technique changes alone produced very significant increase in performance during training.

Most elite swimmers would rejoice at improving performance by more than 3% as a result of one or more year's training. That this effect was achieved in less than a week underscores two things. First, very significant improvements can result from technique alterations, particularly those actions which occur underwater. Second, the application of an effective teaching strategy can produce very rapid behavior and performance changes, the rate of which is rarely witnessed in swimming settings.

The use of continuous contingent reinforcement in the forms of verbal praise and performance information, the consistent prompting of step elements, the significant frequency of correct skill attempts, and the presentation of a progression of learning that facilitated easy learning were part of the "instructional package" in this demonstration. These ingredients should be consistent features of effective coaching. Unfortunately, they are not emphasized enough in coaching manuals, instructional courses, or symposia.

The package of reverse progression teaching elements employed in this investigation offers a clearly defined and measurable strategy for coaching efficient sporting techniques. It produced very few errors in the learning experience and obvious behavior changes. The speed with which the technique changes were achieved was very impressive.

This study focused on and reinforced the importance and practicality of reverse progression teaching as an effective instructional strategy for elite athletes. It was demonstrated as a vehicle for altering established techniques. It is equally effective for producing new techniques but the demonstration of that use will have to remain the province of another investigation.


  1. Ashy, M. H., Landin, D. K., & Lee, A. M. (1988). Relationship of practice using correct technique to achievement in a motor skill. Journal of Teaching in Physical Education, 7, 115-120.
  2. Chelladurai, P., & Stothart, C. (1978). Backward chaining: A method of teaching motor skills. CAHPER Journal, 1, 26-29, 36-37.
  3. Cox, J. A., & Boren, L. M. (1965). A study of backward chaining. Journal of Educational Psychology, 56, 270-274.
  4. Dusault, C. (1986). A backward shaping progression of the volleyball spike approach and jump. Volleyball Technical Journal, 8, 33-41.
  5. Koop, S. & Martin, G. 1. (1983). Evaluation of a coaching strategy to reduce swimming stroke errors with beginning age-group swimmers. Journal of Applied Behavior Analysis, 16, 447-460.
  6. Maglischo, E. W. (1982). Swimming faster. Palo Alto, CA: Mayfield.
  7. Martin, G . L., Koop, S., Turner, G., & Hanel, F. (1981). Backward chaining versus total task presentation to teach assembly tasks to severely retarded persons. Behavior Research of Severe Developmental Disabilities, 2, 117-137.
  8. Rushall, B. S., & Ford, D. (1982). Teaching backwards - an alternative skill instruction progression. CAHPER Journal, 48(5), 16-20.
  9. Simek, T. C., & O'Brien, R. M. (1981). Total golf. A behavioral approach to lowering your score and getting more out of your game. New York, NY: Doubleday.
  10. Spooner, F., & Spooner, D. (1984). A review of chaining techniques: Implications for further research and practice. Education and Training of the Mentally Retarded, 19, 114-124.
  11. Spooner, F., Spooner, D.,& Ulicny, G. (1986). Comparisons of modified backward chaining: Backward chaining with leapaheads and reverse chaining with leapaheads. Education and Treatment of Children, 9, 122-134.
  12. Weiss, K. M. (1978). A comparison of forward and backward procedures for the acquisition of response chains in humans. Journal of the Experimental Analysis of Behavior, 29, 255-259.

Return to Table of Contents for this issue.