Effects of live, videotaped, or written instruction on learning an upper-extremity exercise program.

In today's health care environment, cost containment is important. Efforts to maintain high-quality patient care while controlling costs have led to the use of various modes of instruction that do not require the physical therapist to be present throughout the instructional period. (1) Physical therapy intervention usually includes live one-on-one instruction and may include the use of videotapes or handouts to augment this instruction. Videotapes can be used in clinic or hospital settings to instruct large numbers of patients on the same series of exercises. Handouts can be provided in place of live therapist instruction. Although the videotape and handout modes of instruction can appear to be cost-effective and time-efficient, the suitability of these materials as alternatives to live instruction currently is unknown.

Physical therapists have found themselves struggling to meet the demands of patient care in a changing health care environment. Most therapists appear to be continually striving to provide effective interventions in the brief time allotted. Physical therapy outcomes often rely on the patient's ability to learn the motor skills that constitute an exercise program. As physical therapists are challenged to provide high-quality patient care with fewer personal interactions, alternatives to live instruction become increasingly important. The motor learning literature provides information concerning feedback, modeling, and other issues that may be helpful to physical therapists in maximizing the efficiency of the teaching-learning process.

Feedback is regarded as critical to the learning of motor skills. (2-4) Researchers have identified the type of information feedback that is best for learning skills that involve multiple degrees of freedom, such as the whole-body actions used frequently in the clinical setting. Kernodle and Carlton (5) examined a throwing task using the non-dominant arm. The subjects were provided with one of the following: feedback about the distance thrown, videotaped replay of the action just performed, videotaped replay plus a cue on which to focus their attention, or videotaped replay with transitional information on how to improve the next throw. Subjects practiced this skill over 600 trials and over a 4-week period and viewed a videotaped model after every 10th trial. The researchers found that feedback about the distance thrown or performance of previous throws was less beneficial than was transitional information about the correction to be made and cues about how to focus on the important aspects of the movement.

In a clinical setting, feedback is usually provided in person, and it incorporates cues and guiding information. Less is known about feedback that is provided by an instructor who is not physically present. Riolo (6) examined this question in a study of correct and incorrect modeling during a sterile hand-washing task. Subjects were 60 female physical therapist students who were assigned to 1 of 4 groups: (1) a skilled instruction group that viewed only a videotape of a skilled, error-free model; (2) an error instruction group that observed the error-free model in addition to watching a second section of videotape that demonstrated incorrect and then corrected hand washing; (3) an enhanced skilled instruction group that observed the error-free model in addition to watching a second section of videotape that demonstrated correct hand washing; and (4) a control group that received no videotaped demonstration. For groups 2 and 3, demonstrations in the second section of videotape were labeled on the screen as correct or incorrect.

Research results have indicated positive effects on learning for subjects who observe a live model and then receive feedback about the model's movement outcome. (7,8) Subjects who observe a model who is learning a task may be able to improve their problem-solving skills and perform better on a specified skill. (9,10) Riolo (6) expected that observing videotaped models performing incorrect and then corrected hand washing would improve retention scores by creating a situation in which subjects could learn not only how to perform the correct movements but also how to avoid common mistakes. Contrary to these expectations, no differences were found among the first 3 groups (skilled instruction, error instruction, or enhanced skill instruction). Riolo (6) attributed these results to various limitations of the study, including the simplicity of the task, the short duration of instruction and modeled behavior (observing the task only 2 times), the limited number of practice trials (2) and modeled errors (5 of 17), and a possible ceiling effect on scores.

Although modeling is used frequently in the clinical setting, modeling effects are not as well understood as feedback effects in motor learning. According to social cognitive theory, effective modeling depends on the development of a cognitive representation of the modeled action before the performance of that action. (11) Based on this theory, modeling should occur before performance is required, and the development of a more precise cognitive representation should result in better performance. Carroll and Bandura (12) tested 56 college-aged volunteers to determine their ability to reproduce a modeled demonstration of a complex paddle movement. Subjects were tested on their ability to recognize and sequence pictures of the modeled task as a measure of their cognitive representation of the task. Subjects then were tested on their ability to accurately reproduce the modeled task. Subjects who observed a videotaped model 8 times achieved higher reproduction accuracy and cognitive representation scores than subjects who observed the videotaped model only twice. Subjects who also received verbal descriptions of each motor component as it was performed, along with 8 observations of the model, bad the highest reproduction accuracy and cognitive representation scores.

Research (13) suggests that modeling can be most effective if it is presented at appropriate times during skill acquisition. Learning can be most effective when modeling occurs before practice and is repeated at various intervals during practice. (13) In theory, modeling interspersed with practice may reinforce a cognitive picture while also providing problem-solving opportunities through physical practice.

Although research may provide valuable information that physical therapists can use to instruct patients, many important questions remain unanswered. Feedback in terms of the nature of errors and how to correct those errors seems important. It is unclear, however, if that information needs to be individualized and, therefore, presented by a live instructor. Perhaps more generalized feedback presented in a videotape can produce similar success to live instruction in learning the task. Furthermore, little is known about the relative benefit of modeling provided to the learner directly by live or videotaped instruction.

The purpose of our study was to determine whether different modes of instruction affect the learning of an exercise program as measured by a test of retention of performance skills immediately after instruction and practice and after a 1-day delay. We expected that there would be a main effect of modes of instruction on test scores. Beyond this overall effect, we hypothesized that: (1) test scores would be higher for subjects who received modeling presented by a live instructor or videotape as compared with subjects who received written instruction only, (2) test scores would be higher for subjects who received live instruction with individualized feedback as compared with subjects who received videotaped or written instruction, and (3) test scores would be higher for subjects who received videotaped instruction that showed common errors and information about correcting those errors as compared with subjects who received videotaped instruction with correct modeling demonstrations only.

Method

Subjects

Forty-three volunteer subjects who were 26 to 51 years of age ([X.bar]=38.4, SD=7.4) were recruited for participation in the study. Data for 3 subjects were excluded from analysis because they did not adhere to study requirements concerning practice time. Characteristics of the remaining 40 subjects who participated in the study are detailed in Table 1. Twenty subjects participated from each of 2 age groups: 25 to 40 years and 41 to 55 years. Although age was not a variable in this study, subjects were grouped by age in order to obtain an equal sample that was representative of the general adult population and to control for any possible age effects on motor learning or performance. (14) Subjects completed a questionnaire to report age, sex, highest education level completed, race and ethnicity, general health, and frequency and type of exercise. Inclusion criteria were as follows: 25 to 55 years of age, completion of a high school education, English speaking, no known neurological disorder, no recent upper-quarter musculoskeletal injury that affected daily activity for more than 2 days in the 6 weeks prior to the study, adequate pain-free range of motion and muscle force for performance of the exercise program, and normal or corrected normal vision and hearing by self-report. Subjects with a self-reported history of receiving formal physical therapy training or patient education as part of a physical therapy intervention for the upper extremity also were excluded.

Subjects in each age group were randomly assigned to 1 of 4 groups that differed in the type of instruction received: (1) a live skilled model, (2) a videotape of a skilled instructor with a model who demonstrated errors and correction, (3) a skilled model in an error-free videotape, and (4) instruction by handout only.

Instructional Materials

Subjects were instructed in a series of 5 exercises. The exercises were selected because they are commonly used in physical therapy upper-quarter rehabilitation programs, they could be scored from a videotape, and they could be performed in a standing position without special equipment other than a simple resistive band. A very light resistive band, yellow Thera-Band,* was chosen for use in the study The 5 exercises, which subjects performed in the standing position with a 90-cm (3-ft) piece of yellow Thera-Band, were: (1) bilateral scapular retraction or "both arms 'W' shoulder blade squeeze" (Fig. 1A), (2) unilateral elbow extension or "one arm elbow straightening" (Fig. 1B), (3) unilateral shoulder flexion in scapular plane or "one arm 'full' soda can exercise" (Fig. 1C), (4) bilateral shoulder circles or "shoulder clocks" (Fig. 1D), and (5) bilateral shoulder flexion or "double arm 'V' exercise" (performed without the resistive band) (Fig. 1E).

The exercises were chosen with the assistance of 4 experienced physical therapists (mean years of clinical experience=10, SD=6.8, range=5-20). These therapists also helped to identify 4 critical components of each exercise, which provided the basis for scoring and analysis of exercise performance. These critical components were based on elements that are often performed incorrectly by patients but are deemed essential for correct completion of each exercise. Critical components also were selected to allow scoring of each component from videotapes. Certain components that are commonly performed incorrectly in the clinical setting were selected for demonstration and correction on the videotape with corrected errors. Table 2 provides details of the critical components as well as additional components used for scoring.

Instruction provided to all subjects included an introduction that stated the importance of using the proper technique lot performing each of the 5 exercises. Subjects were instructed in the use of the Thera-Band, including correct hand placement on the band, how to maintain tension, and how to control the entire movement without the quick recoiling of the band. Subjects also were advised to remember the name of each exercise, because only the names would be provided to them during testing.

Two videotapes, a script for live instruction, and a handout were used in the study. The model in the videotapes and the live instruction was the primary investigator (JAR), who is a physical therapist who had 8 years of clinical experience at the time of the study. To maintain consistency with verbal instructions and modeling in all groups, the primary investigator referred to a paper with cues (the critical components) for the demonstration of each exercise. In an effort to ensure consistency, a checklist of events was used dining each instruction and testing procedure.

All videotaped and live instruction began with exercise 1, "both arms 'W' shoulder blade squeeze" (Fig. 1A). The videotaped or live model provided a preview of the exercise, a period of more detailed instruction on the exercise, and a review of the exercise. Consistent with past research results, (12,15) frequent modeling demonstrations were incorporated into the videotaped and live instruction. Modeling was provided once during preview, 5 times during instruction, and once again during review for a total of 7 demonstrations. Time for 5 consecutive practice trials was provided after instruction of each exercise. This sequence was repeated for each of the remaining 4 exercises. Time allotted for the introduction, instruction, and practice time was 9 minutes.

[FIGURE 1 OMITTED]

In the live instruction as well as in the error-tree videotape, the model correctly instructed and modeled each component of each exercise. In the corrected-error videotape, the model instructed a male novice who made 10 different errors, 2 for each exercise, during the course of the videotape (see Tab. 2 for a description of the errors the novice demonstrated). During the second axed fourth demonstrations of each exercise, the model provided verbal feedback to the novice about the error he made, and the novice corrected the error on his next attempt at the exercise. Verbal feedback included specific statements or cues about how to correct the component error, such as "be sure to keep your elbows straight the whole time" or "keep your thumb pointed up like you are holding a full can of soda."

Subjects in the handout group received instruction from a written handout with pictures that illustrated each exercise axed with text that defined the 4 critical components of each exercise. Figure 1 replicates the handout provided to each subject, but differs from the actual handout in font size and size of the illustrations. The handout used in the study was a total of 3 pages, with each exercise illustration and text covering half a page. The quality of the handout was designed to be comparable to that of handouts commonly provided to patients in physical therapy clinical settings. We consulted 3 therapists, who were chosen because they served as advisors on the thesis committee, about the nature of the pictorial representation axed the wording of the text accompanying each exercise. We conducted pilot testing of the handout with one subject who had not received previous instruction about the upper-extremity exercises. Based on the input of the aforementioned individuals, inset pictures were added and clarification was made to the wording of the text.

Procedure

Instruction. Each of the subjects who participated in the study did so on 2 separate days. On the first day, procedures were explained, and the subject was screened for acceptable shoulder and elbow range of motion and muscle force. All subjects signed an informed consent form approved by the Committee on the Protection of Rights of Human Subjects at the University of North Carolina at Chapel Hill. The subject was then randomly assigned to 1 of 4 groups by drawing slips of paper grouped by age.

Research on context (16,17) suggests that errors may occur when people are instructed in one context and then tested in a dissimilar context. To try to maintain consistency of context ill our study, instruction and testing took place in the same location for any given subject. The majority of the subjects (35/40) were tested in their homes. Three subjects were tested in the researcher's home: 2 because of videocassette recorder (VCR) failure at the subject's home and 1 for subject convenience. Two subjects were tested in a small workplace library. Background noise and distractions were minimized by using a closed or protected space. All subjects wore short-sleeved shirts to allow the researcher to visualize the elbow and arm for scoring. A video camera for recording each subject's activities and movements during instruction was placed to the side of the subject as he or she stood facing the medium of instruction.

Subjects in the live instruction group received error-free exercise instruction, meaning the model demonstrated the correct form of each exercise. Because the subjects in this group had live and not videotaped instruction, they received live individualized verbal feedback about any errors in their performance. Verbal feedback was provided to each subject after practice trials 2 and 4 of each exercise. Verbal feedback may have included statements about how to correct errors in critical components, such as "be sure to keep your elbows straight the whole time" or "hold the hand taut between your bands." if all components were appropriate, verbal feedback confirmed this with the statement "that is correct." The video camera was started at the beginning of the instructional session. Instruction and modeling were provided before the practice time for each exercise, and subjects were advised to not practice on their own before this time. Subjects were given a Thera-Band and asked to perform 5 practice trials of each exercise.

Participants in the corrected-error videotape group and the error-free videotape group individually viewed the appropriate videotape. Each subject was told not to practice along with the videotaped demonstrations because designated practice periods would be provided in the videotape. The subject was given the piece of Thera-Band, and the VCR was started. As with the live instruction group, the video camera was started at the beginning of the instructional session. Additional questions, clarification, or conversation were not allowed. The subject was left alone in the room during the viewing of the videotape. Upon completion of the videotape segment, the primary investigator reentered the room, and the VCR was stopped.

In the handout group, subjects received instruction only from the written handout. Additional questions, clarification, or conversation were not allowed. Each subject was given a Thera-Band and instructed in its use, and the subject was instructed to read and then practice each exercise 5 times, beginning with the first exercise (Fig. 1A). The video camera was started immediately after the subject received the handout and instructions. The subject was left alone in the room and given 9 minutes (consistent with the instruction time of the other groups) to review the handout and perform the practice trials. At the end of the 9-minute period, or when the subject indicated that he or she was ready, the primary investigator reentered the room. Before testing, the handout was returned to the primary investigator.

Testing. Subjects in all 4 groups were tested immediately after the completion of the instruction and practice period and again 24 hours ([+ or -] 6 hours) after the instruction and practice period. The 24-hour retention interval has been shown to be adequate time for motor memory consolidation. (18) Subjects were given a "cue board," which stated only the name of each exercise. Subjects were not allowed to refer to the written handout during testing. The "cue board," aligned vertically and ordered the same as the instructions, read as follows: "'W' shoulder blade squeeze"; "Elbow straightening"; "'Full' soda can exercise"; "'Shoulder clocks'"; and "Double arm 'V' exercise." This cue board was provided to trigger the subjects' memory of the exercises but not to provide any additional instructions or pictures.

During all testing, subjects were videotaped in a position in which they faced the camera at a 45-degree angle (aligned with a tape mark on the floor). The primary investigator stood behind the video camera during testing. Subjects performed the exercises in any order desired, completing 5 repetitions of each exercise. To minimize any effects associated with beginning or ending the performance effort, only the middle 3 repetitions were used for scoring purposes. Subjects were free to pause for thought or take their time during testing. Time restrictions did not exist. At the completion of testing, subjects were asked to return the Thera-Band. In order to control for practice time across the groups, subjects were instructed not to practice any of the exercises until the next test session. Subjects also were instructed not to discuss the exercises with anyone during the interval between tests.

Subjects were tested for retention after approximately 24 hours. They were provided no additional modeling in written, videographic, or verbal form. Subjects were given the Thera-Band and the cue board, and they were instructed to perform each exercise 5 times. At this point, the video camera was started. As with the immediate posttest scoring, only the middle 3 repetitions were used for scoring purposes.

Practice time was controlled in each group. Data from 3 subjects had to be excluded because review of the videotapes revealed that these subjects practiced more than 5 repetitions of each exercise. After completion of retention testing, subjects were asked whether they had practiced or discussed any of the exercises during the 24-hour period between tests. No subjects reported practicing or discussing the exercises during the testing interval.

Data Analysis

All videotapes were reviewed to ensure that subjects performed 5 practice trials of each exercise during the instructional period. The videotapes of subject performance were reviewed and scored by the primary investigator using the checklist of critical components (Tab. 2). In addition, review of videotapes allowed the primary investigator to score errors during the instruction period.

Subjects received 1 point for each exercise component that was performed correctly, with a maximum score of 20 for the series of 5 exercises. With the scoring of 3 repetitions of each exercise, this provided a total of 60 maximum points for each subject.

Intrarater reliability was determined by repeated scoring, 1 week apart, of the videotaped performances of 8 subjects from the testing period immediately after the instruction and practice session. These subjects were the first 2 subjects tested in each of the 4 instructional groups. Interrater reliability was determined by asking a physical therapist with 7 years of clinical experience, who was unaware of group assignment of the subjects, to independently score the same 8 videotaped performances. An intraclass correlation coefficient (ICC [3, 1]) (19) was used for analysis of intrarater and interrater reliability. The ICC values were .98 for intratester reliability and .95 for intertester reliability. Because the ICC is more sensitive to random error than to systematic error, (20) the F ratio from the repeated-measures analysis of variance (ANOVA) was used to assess the significance of differences between repeated scoring of the videotaped performances. The F ratios were not significant (P>.05), indicating no systematic differences in scoring within or between raters.

Statistical analysis consisted of a 2-way (4 x 2) analysis of covariance (ANCOVA) with repeated measures on the second factor and age as a covariate. (21) Age was used as the covariate because of possible changes in motor learning abilities during middle age. Following a significant omnibus F test, Tukey honestly significant difference (HSD) post hoc analyses were performed to identify differences among groups. All statistical analyses were performed with a significance level of P<.05.

Results

Results of the performance accuracy scores of all subjects, by group and time of testing, are displayed in Figure 2. Results of the repeated measures ANCOVA revealed that there was a group effect (F=8.01; df=3,35; P<.001), but neither the time of testing nor the group x time interaction was significant. Tukey post hoc comparisons indicated that the handout group was less accurate than the other 3 groups (HS[D.sub..05]=4.74). The live instruction group and the 2 videotaped instruction groups did not differ from one another in performance accuracy.

Total errors committed by each group on each component of the 5 exercises are summarized in Table 2. Total errors represent the combined scores from immediate and retention testing. The live instruction group and the videotaped instruction groups averaged 148.3 total errors compared with 345 total errors for the handout group. Fifteen percent of total errors for all exercises involved incorrect use of the Thera-Band (incorrect placement or inability to maintain tension throughout the range of motion). The combined accuracy of all groups for each of the 5 exercises varied from 80% to 86%.

The subjects were videotaped while practicing during the exercise instruction portion of the study. Although these videotapes were not scored for inclusion in the statistical analyses, they were reviewed and analyzed. Upon review, it was noted that almost all subjects in the videotaped instruction groups were able to correctly practice the exercise just modeled. Additionally, the live instruction group required very little extrinsic feedback to achieve correct performance during the practice time of the instructional session. In contrast, the practice videotapes of the handout group revealed that 8 of the 10 subjects performed incorrect movements during practice of at least 1 of the 5 exercises. Three of these subjects performed only 1 exercise incorrectly; the remaining 5 subjects performed 2 of the 5 exercises incorrectly during practice. The exercises performed incorrectly varied by individual.

No subjects reported performing actual practice trials between sessions, although some reported thinking about the exercises and mentally rehearsing. Subjects in all groups often had comments during or after testing, usually related to their difficulty in remembering the exercises. Comments about remembering the exercises usually included confusion about different components of the exercises (eg, "I know my thumbs are pointed up for at least one of these exercises"). Although several subjects struggled with memory recall of some components of the exercises, only one subject was unable to complete an entire exercise because of a reported inability to recall that exercise or exercise cue.

Discussion

Our results indicate that mode of instruction affects learning of an exercise program as measured by tests of performance accuracy immediately after instruction and practice and after a 1-day delay. Our hypothesis, that test scores would be higher for subjects who received modeling presented by a live instructor or videotape compared with written instruction only, was supported. Subjects in the handout group were less accurate and had a greater number of errors than subjects who received live or videotaped modeling and exercise instruction. Our findings support data that suggest that modeling enhances the quality of performance (15,22) and is beneficial for learning a new motor skill. (15,23) The findings further support research suggesting that use of brochures or handouts alone results in poorer exercise performance compared with videotaped instruction (24) or therapist-based instruction. (25)

Our second hypothesis, that live instruction with individualized feedback would result in better performance compared with videotaped or written instruction, was only partially supported. Although performance accuracy scores for subjects who received live instruction were higher than those for subjects who received written instruction, live instruction did not produce higher scores than videotaped instruction. Furthermore, the 2 groups that received videotaped instruction did not differ in the quality of performance. This result indicates a lack of support for our third hypothesis, that the addition of common errors and information about correction of those errors would enhance learning compared with videotaped instruction with only correct modeling demonstrations.

The better performance for subjects receiving live or videotaped instruction compared with those given only written instruction may have resulted from development of a more precise cognitive representation of the required movement patterns by these subjects. (12,26) Live and videotaped modeling may have allowed for an adequate cognitive representation, whereas the pictures in the handout may not have given subjects sufficient information about the movement components of each exercise. Even with equal physical practice by each group, the lack of observational (ie, modeled) information led to poorer performance by the handout group. This supports past research that modeling may be as effective or more effective than physical practice. (23)

People in the handout group may not have understood the proper exercise form. This may have resulted in poorer performance and the higher frequency of incorrect practice trials among subjects in this group. Although subjects in all groups forgot some components of exercises as noted during testing, in our opinion, subjects in the handout group often appeared to have an incomplete understanding of the exercises throughout the instructional session and practice interval. Use of a handout alone resulted in more than double the number of performance errors associated with other types of instruction.

Physical therapists prescribe exercises usually to achieve specific muscle or muscle group performance and to minimize injury. If an exercise is performed incorrectly, then isolated strengthening of the targeted muscle or muscle group is unlikely to occur. In some instances, incorrect exercise performance can result in additional injury to the muscle or surrounding tissue. (27) For example, if the double arm "V" exercise was performed with the thumbs pointing down instead of up, elevation beyond 90 degrees could place the person at risk for developing subacromial impingement. (28) To the extent that correct performance of all prescribed exercises is important for achieving the desired benefits, a 100% increase in errors by the handout group compared with the other groups could reflect a meaningful effect. Results of the present study, in combination with those by Weeks at al (24) and Friedrich et al, (25) suggest that receiving a handout alone may not be sufficient for learning a basic exercise program.

Because feedback is regarded as critical to the learning of a motor skill, (2-4) the lack of feedback provided to the handout alone group also may have played an important role in their poorer performance. The results of our study, however, cannot be fully explained by differences in the feedback given to all of the 4 groups. Despite receiving different types of feedback, the videotaped instruction groups did not differ from one another or from the live instruction group in performance accuracy.

In contrast to the more specific temporal or spatial tasks employed in other research (eg, rapid elbow flexion to reach a specific angle), (2,3,29) we used what we considered clinically relevant shoulder exercises and scored them on the basis of completion of 4 critical components. Additional differences among the modes of instruction might have been observed if a more complex program of exercises had been used. Although chosen for their clinical relevance, the exercises we used involve movements that are relatively common during daily activities. These movements would be categorized as closed skills, or skills that are performed in a predictable environment. (30) The results of this study apply to exercise programs involving similar movements and should not be generalized to exercises that would be categorized as open skills, or skills that are performed in a constantly changing environment. (30,31)

Performance scores for this study were generally high, indicating that the exercises were relatively easy for these subjects to perform. Subjects were not asked to learn unusual movement patterns (eg, pelvic tilts or Kegel exercises) that might have challenged their motor learning abilities to a greater extent. In addition, subjects were exposed to only 2 errors on the corrected-error videotape. A more challenging task with more modeled errors might have led to greater differences among some of the groups.

Differences between live and videotaped instruction groups also might have been observed if adults with injuries or lower levels of education were used as subjects. The subjects were a fairly homogenous group, all were well educated, English speaking, and generally active. Subjects had no injuries and may have been motivated by their involvement in a research study and the use of a video camera rather than by a need to engage in a rehabilitation program. A patient population might have increased the clinical relevance of the results. Subjects without injuries, however, were chosen in order to control for factors that could have altered movement form (and therefore scoring criteria) such as pain avoidance patterns, muscle weakness or substitution, or limited endurance. The lower accuracy of the subjects in the handout group suggests that a general patient population might have similar, or even more profound, difficulties with this mode of instruction.

Our study had several limitations that should be taken into account when interpreting the results. The primary investigator instructed and scored the subjects and therefore was not blinded to group selection. The high intertester reliability, however, suggests to us that experimenter bias was not a major issue. Another limitation relates to the quality of the written handout materials used in the study. Performance accuracy scores of the handout group may have reflected misinterpretation of the wording or diagrams used in the handout. The handout was developed to reflect those commonly used in clinical settings and to include illustrations along with written text in order to improve understanding. (32) A handout with wording and diagrams that have been studied for clarity might have resulted in better performance for the handout group. A third limitation relates to the time allowed for memory recall during testing. Although time of instruction and number of practice repetitions were controlled, the amount of time permitted for performance of the exercises during testing was not controlled. Some subjects required more time to recall the exercises than others. We did not attempt to control for mental rehearsal during testing; instead we sought to allow subjects adequate time to recall the exercises and produce their best performance of each exercise.

We believe that the focus of our research was on examining different modes of instruction. If exercises are not appropriate for a patient's needs, then the mode of instruction is irrelevant. Development of an effective exercise program requires an examination and assessment of the patient's specific impairments and functional limitations. Once appropriate exercises have been identified, then audiovisual materials may help the patient perform the exercises correctly.

Most subjects in our study demonstrated some errors during testing. We believe it is useful to know that this occurs. Regardless of the type of instruction, we contend that people should be observed demonstrating previously taught exercises to ensure correct performance.

Conclusion

Live and videotaped modeling are more effective than a handout alone for achieving accuracy in the performance of a basic exercise program by adults without injuries who have a fairly high level of education. Videotaped instruction can be an economical choice in some clinics or hospitals, and this may be appropriate for some patients and for some types of exercises. Further research is needed to determine whether the mode of instruction affects motor skill learning in groups of patients who are injured or otherwise compromised or for a more unfamiliar or complex skill.

Table 1.
Characteristics of Subjects by Group

                         Sex              Age (y)

Group              M      F    [bar.X]    SD    Range

Live modeling      6      4    37.7       7.9   29-50
Corrected-error    4      6    38.6       5.6   29-44
  videotape
Error-free         3      7    37.7       8.3   26-51
  videotape
Handout            4      6    39.4       7.7   29-50

                  Race/Ethnicity (a)     Education (b)    Exercise (c)

Group              W      H      I     HS     CO     PS    Yes    No

Live modeling     10      0      0      1      7      2     9      1
Corrected-error    9      0      1      0      5      5     7      3
  videotape
Error-free         8      1      1      1      4      5     8      2
  videotape
Handout           10      0      0      0      3      7     8      2

(a) W=white, H=Hispanic, I=Indian (Asian).

(b) HS=high school, CO=college, PS=post-secondary.

(c) Yes=exercise at least once per week, No=no regular exercise.

Table 2.
Description of Components Used for Scoring Each Exercise
With Number of Errors by Component and Group (a)

                                                      No. of Errors
                                                         by Group

Exercise              Scoring Components              1    2    3    4

1. Both arms "W"      Hands in front less than        2    4   16   23
   shoulder blade       shoulder width apart
   squeeze            Elbows bent the same angle      0   11   22   33
                        throughout exercise (b)
                      Scapular retraction             0   12   14   23
                        (elbows extend beyond
                        frontal plane) (b)
                      Band snug without slack or      5    2   11   13
                        snapping
2. One arm elbow      Both arms begin with            5   21   17   22
   straightening        elbows bent and hands at
                        respective shoulders (b)
                      Anchor arm remains bent in      0    9    6    6
                        same position
                      Dominant arm extends to         6   27   12   12
                        side (<45[degrees] of
                        abduction) (b)
                      Band snug without slack or      3    6    4   13
                        snapping
3. One arm "full"     Thumb begins and remains        0    0    0    4
   soda can exercise    pointed up (b)
                      Arm raised only to shoulder    12    0   10   30
                        level (<120[degrees] of
                        flexion)
                      Elbow begins and remains       14    0   14   28
                        straight (b)
                      Band under one foot without     6   18    9    7
                        slack
4. "Shoulder clocks"  Elbows remain straight the     16   13    9   13
                        entire time
                      Rotation occurs only at the     5    4    0    6
                        shoulders (b)
                      Rotation occurs, beginning      6    0    3   36
                        forward (b)
                      Band is secured under both      3    6    0   12
                        feet without slack or
                        snapping
5. Double arm "V"     Elbows begin and remain         3    0   17    6
   exercise             straight the whole
                        time (b)
                      Arms raise together overhead    6    9    9   19
                        to "V" position (b)
                      Arms in scapular plane         21    5    9   21
                      Thumbs pointed up the whole     0    3    0   18
                        time
Totals                                              113  150  182  345

(a) Group 1=live modeling, group 2=corrected-error videotape
group, group 3=error-free videotape group, group 4=handout group.

(b) Critical component used for error demonstration in
video for corrected-error videotape group (group 2).

References

(1) Meredith P, Emberton M, Wood C. New directions in information for patients. BMJ. 1995;311:4-5.

(2) Winstein CJ, Pohl PS, Lewthwaite R. Effects of physical guidance and knowledge of results on motor learning: support for the guidance hypothesis. Res Q Exerc Sport. 1994;65:316-323.

(3) Winstein CJ, Pohl PS, Cardinale C, et al. Learning a partial-weight-bearing skill: effectiveness of two forms of feedback. Phys Ther; 1996; 76:985-993.

(4) Salmoni AW, Schmidt RA, Walter CB. Knowledge of results and motor learning: a review and critical reappraisal. Psychol Bull. 1984;95: 355-386.

(5) Kernodle MW, Carlton LG. Information feedback and the learning of multiple-degree-of-freedom activities. J Mot Behav. 1992;24:187-196.

(6) Riolo L. Effects of modeling errors on the acquisition and retention of sterile hand washing task. Percept Mot Skills. 1997;84:19-26.

(7) McCullagh P, Caird JK. Correct and learning models and the use of model knowledge of results in acquisition and retention of a motor skill. Journal of Human Movement Studies. 1990;18:107-116.

(8) Adams JA. Use of the model's knowledge of results to increase the observer's performance. Journal of Human Movement Studies. 1986;12: 89-98.

(9) Lee TD, Swanson LR, Hall AL. What is repeated in a repetition? Effects of practice conditions on motor skill acquisition. Phys Ther. 1991;71:150-156.

(10) Hodges NJ, Franks IM. Modeling coaching practice: the role of instruction and demonstration. J Sports Sci. 2002;20:793-811.

(11) Carroll WR, Bandura A. Role of timing of visual monitoring and motor rehearsal in observational learning of action patterns. J Mot Behav. 1985;17:269-281.

(12) Carroll WR, Bandura A. Representational guidance of action production in observational learning: a causal analysis. J Mot Behav. 1990;22:85-97.

(13) Weeks DL, Anderson LP. The interaction of observational learning with overt practice: effects on motor skill learning. Acta Psychol (Amst). 2000;104:259-271.

(14) Feltz DL. The effects of age and number of demonstrations on modeling of from and performance. Res Q Exerc Sport. 1982;53:291-296.

(15) Sidaway B, Hand MJ. Frequency of modeling effects on the acquisition and retention of a motor skill. Res Q Exerc Sport. 1993;64:122-126.

(16) Wright DK, Shea CH. Cognition and motor skill acquisition: contextual dependencies. In: Reynolds CR, ed. Cognitive Assessment: A Multidisciplinary Perspective. New York, NY: Plentium Press; 1994: 89-106.

(17) Gomez GE, Gomez EA. Learning of psychomotor skills: laboratory versus patient care setting. J Nurs Educ. 1987;26:20-24.

(18) Walker MP, Brakefield T, Hobson A, Stickgold R. Dissociable stages of human memory consolidation and reconsolidation. Nature. 2003; 425:616-620.

(19) Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:429-428.

(20) Norton BJ, Ellis JB. Reliability and concurrent validity of the Metrecom for length measurements on inanimate objects. Phys Ther. 1993;73:266-274.

(21) Blalock HM Jr, ed. Social Statistics. 2nd ed. New York, NY: McGraw-Hill; 1979.

(22) McCullagh P, Stiehl J, Weiss MR. Developmental modeling effects on the quantitative and qualitative aspects of motor performance. Res Q Exerc Sport. 1990;61:344-350.

(23) Shea CH, Wright DL, Wulf G, Whitacre C. Physical and observational practice afford unique learning opportunities. J Mot Behav. 2000;32:27-36.

(24) Weeks DL, Brubaker J, Byrt J, et al. Videotape instruction versus illustrations for influencing quality of performance, motivation, and confidence to perform simple and complex exercises in healthy subjects. Physiotherapy Theory and Practice. 2002;18:65-73.

(25) Friedrich M, Cermak T, Maderbacher P. The effect of brochure use versus therapist teaching on patients performing therapeutic exercise and on changes in impairment status. Phys Ther. 1996;76:1082-1088.

(26) Bandura A. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice-Hall; 1986.

(27) Cools AM, Witvrouw EE, Declercq GA, et al. Scapular muscle recruitment patterns: trapezius muscle latency with and without impingement symptoms. Am J Sports Med. 2003;31:542-549.

(28) Townsend H, Jobe FW, Pink M, Perry J. Electromyographic analysis of the glenohumeral muscles during a baseball rehabilitation program. Am J Sports Med. 1991;19:264-272.

(29) Vander Linden DW, Cauraugh JH, Greene TA. The effect of frequency of kinetic feedback on learning an isometric force production task in nondisabled subjects. Phys Ther. 1993;73:79-87.

(30) Schmidt RA. Motor Control and Learning: A Behavioral Emphasis. 2nd ed. Champaign, Ill: Human Kinetics Publishers Inc; 1988:578.

(31) Weeks DL. A comparison of modeling modalities in the observational learning of an externally paced skill. Res Q Exert Sport. 1992;63: 373-380.

(32) Michielutte R, Bahnson J, Dignan MB, Schroeder EM. The use of illustrations and narrative text style to improve readability of a health education brochure. J Cancer Educ. 1992;7:251-260.

JA Red, PT, MS, was a student at the University of North Carolina at Chapel Hill at the time this research was completed in partial fulfillment of the requirements for her Master of Science degree in Human Movement Science. Address all correspondence to Ms Red at 719 Clarion Dr, Durham, NC 27705 (USA) ([e-mail]).

VS Mercer, PT, PhD, is Assistant Professor, Division of Physical Therapy, Department of Allied Health Sciences, University of North Carolina at Chapel Hill, Chapel Hill, NC.

Both authors provided concept/idea/research design, writing, and data analysis. Ms Red provided data collection, subjects, and facilities/ equipment. Dr Mercer provided consultation (including review of manuscript before submission). The authors thank Deborah Thorpe, PT, PhD, PCS, and Michael Gross, PT, PhD. fur contributing to concept/idea/research design and consultation (including review of manuscript before submission).

This study was approved by the Committee on the Protection of the Rights of Human Subjects at the University of North Carolina at Chapel Hill.

This article was submitted March 14, 2003, and was accepted January 7, 2004.