This article was downloaded by: [7.197.66.52] On: 13 July 214, At: 12:56 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 172954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Research Quarterly for Exercise and Sport Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/urqe2 Effects of Physical Guidance and Knowledge of Results on Motor Learning: Support for the Guidance Hypothesis Carolee J. Winstein PhD a, Patricia S. Pohl a & Rebecca Lewthwaite b a Department of Biokinesiology and Physical Therapy, University of Southern California b Center for Research in Clinical Biokinesiology, Rancho Los Amigos Medical Center, Downey, California, USA Published online: 8 Feb 213. To cite this article: Carolee J. Winstein PhD, Patricia S. Pohl & Rebecca Lewthwaite (1994) Effects of Physical Guidance and Knowledge of Results on Motor Learning: Support for the Guidance Hypothesis, Research Quarterly for Exercise and Sport, 65:4, 316-323, DOI: 1.18/271367.1994.167635 To link to this article: http://dx.doi.org/1.18/271367.1994.167635 PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content ) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Research Quarterly for Exercise and Sport 1994 bytheamerican Alliance for Health, Physical Education, Recreation and Dance Vol. 65, No.4,pp,316-323 Effects of Physical Guidance and Knowledge of Results on Motor Learning: Support for the Guidance Hypothesis Caro/ee J. Winstein, Patricia S. Poh/, andrebecca Lewthwaite The guidance hypothesis (Schmidt, 1991) predicts that theguiding propfffties ofaugmentedfeedback are beneficialfor motor learning when used to reduce error, but detrimental when relied upon. Therefore, a heavilyguidingform offeedback might be detrimentalfor learning. In addition, the guidance hypothesis predicts that practice with a high relativefrequency ofaugmented feedback would bedetrimentalfor learning. An experiment is described that crossed twoforms offeedback with two levels of relativefrequency. Subjects practiced movements to a target with eithe:physicalguidance or knowledge ofresults, and with either a high orfaded relativefrequency. The high frequency physicalguidance condition resulted in thepoorest retention, and both highfrequency feedback conditions resulted in the leastaccuracy in transfer. These results provide supportfor the guidance hypothesis and suggestconsideration ofthe combined effects on learning ofthe typeand relativefrequency ofaugmentedfeedback and acquisition-test conditions. Key words: learning, guidance, knowledge of results, motor skills It is well known that feedback about performance is one of the most powerful variables affecting the learning ofmotor skills (Magill, 1993). Knowledge of results (KR) refers to verbalizable, externally provided postresponse information about the extent of error or success in achieving the outcome goal. Stimulated by a review of the KR literature (Salmoni, Schmidt, & Walter, 1984), recentwork has focused on the differential performance and learning effects ofpractice under conditions ofseveral KR variations, including KR relative frequency (percentage of trials for which KR is provided), summary KR (provision ofkr after a set of trials called the summary length), and KR delay (provision of KR after some postresponse interval). Together, the findings from these studies suggest that practice in conditions with less frequent and less immediate KR is more detrimental for performance but more beneficial for the learning of motor skills than practice in conditions where KR is provided following every trial and with little Submitted: March 3, 1994 Revision accepted: August 18, 1994 Carolee J. Winstein is anassistant professor andpatricia S.Pohl is a doctoral candidate in thedepartment of Biokinesiology and Physical Therapy at theuniversity ofsouthern California. Rebecca Lewthwaite is director at thecenter for Research in Clinical Biokinesiology of therancho Los Amigos Medical Center in Downey, California. postresponse delay (e.g., Schmidt, Young, Swinnen, & Shapiro, 1989; Swinnen, Schmidt, Nicholson, & Shapiro, 199; Winstein & Schmidt, 199). Parallel findings have been observed for motor learning using a presentation-test paradigm (Hagman, 1983). In this paradigm, adapted from verbal learning research, a presentation trial is one in which movement is made to a physical block located at the target location. In contrast, a test trial is one in which movementis made without the physical block in place. A presentation trial is considered a form of guided practice in which an external physical constraint is provided (Newell, Morris, & Scully, 1985). A test trial is considered a form ofrecall practice in which the performer is required to produce the end position without an external constraint. In three experiments, Hagman (1983) showed that practice under a condition of infrequent presentation trials was more detrimental for performance but more beneficial for motor skill learning than practice under conditions with frequent presentation trials. Similar effects have been observed in motor learning studies in which other forms ofguidance have been used (e.g., Annett & Kay, 1957; Armstrong, 197; Lee, White, & Carnahan, 199, Experiment 1). In both the KR relative frequency and presentationtest paradigms, the provision of augmented feedback (KR and physical block, respectively) directs the subject to the goal and helps establish a reference of correctness. These feedback trials arejuxtaposedwith opportunities (no-kr and test trials) for learner-directed processing operations such as retrieval of the action plan from memory. In contrast, KR trials and presentation 316 ROES: December 1994
Winstein, Pohl, andlewthwaite trials differ in the mode (verbalizable vs. physical), timing (postresponse vs. instantaneous), and quality (direction and magnitude of error referenced to the goal vs. on-target experience) of the augmented information. In addition, KR influences performance on the next trial, whereas a presentation trial influences that trial. Thus, these two paradigms appear to differ in the extent to which the augmented feedback guides performance to the target. That is, the augmented information provided by a presentation trial is considered more guiding than that provided by KR in that it guarantees on-target performance. The guidance hypothesis has been proposed to explain the aforementioned KR findings (Salmoni et al., 1984; Schmidt, 1991; Schmidt et al., 1989). In essence, augmented feedback is thought to have guiding properties that have both beneficial and detrimental effects on motor learning. The beneficial effects are thought to be the well-known informational properties of augmented feedback in which knowledge about outcome is used to correct errors and improve subsequent performance. The effectiveness ofaugmented feedback in error identification and reduction, however, is thought to be detrimental in that it prevents or interferes with critical between-trial information processing involving encoding, storage, and retrieval operations, such as problem solving, known to be important for learning (Bjork, 1988; Landauer & Bjork, 1978; Schmidt, 1991). Thus, the guidance hypothesis proposes that a practice schedule with relatively infrequent augmented feedback trials would be optimal in that it would maximize the beneficial effects and minimize the detrimental effects ofaugmented feedback on motor skill learning. Further, if the guiding properties of augmented feedback interfere with processes important for learning, feedback that is relatively more guiding would be expected to have greater detrimental effects on motor learning. The experiment presented here is the first to formally compare the KR and presentation-test paradigms to determine the relative impact on motor learning of feedback frequency schedules and the two types ofaugmented information represented in these different approaches. If these two types of augmented feedback have differential guiding properties, we hypothesized that there would be an interaction between the type of augmented feedback and the frequency with which it was provided. That is, feedback that is highly guiding (i.e., presentation) when provided frequently would be the most detrimental for motor learning. By comparison, feedback that is relatively less guiding in nature (i.e., KR) could be provided at the same high frequency without having as detrimental an effect on motor learning. In contrast, feedback that is highly guiding when provided infrequently would be the most beneficial for motor learning. By comparison, feedback that is relativelyless guiding in nature when provided infrequently RDES: December 1994 would not be as beneficial for motor learning. The purpose of this experiment was to examine the contribution ofboth type and frequency ofaugmented feedback on the learning ofa target location. Method Subjects Subjects were 4 right-hand dominant graduate student volunteers (31 females, 9 males) from the University ofsouthern California, who were between the ages of22 and 37 (M=25.2 years, SD=3.9). All subjectswere naive to the specific purposes of the experiment and had no prior exposure to the experimental apparatus. Subjects were randomly assigned to one offour practice conditions. Design The experimental design consisted oftwo betweensubject factors offeedback type (presentation, KR) and feedback relative frequency (high, fade) and a withinsubject repeated measures factor of trials or trial block. The two between-subject factors were crossed resulting in four conditions ofpractice: High-Presentation, Fade Presentation, High-KR, and Fade-KR. The acquisition phase consisted of six blocks of six trials for a total of36 trials. Subjects in the High-Presentation practice condition moved to a physical block placed at the target on the first five of every six trials and moved to the subject-recalled target without the block on every sixth trial. Subjects in the Fade-Presentation practice condition moved to the block at the target in a faded schedule, at an overall relative frequency of 33% and an absolute frequency of 12 augmented information trials. Specifically, the proportion ofaugmented feedback trials was 5% in the first two 6-trial blocks,, 33% in the next two 6-trial blocks, and 16% in the last two 6-trial blocks. Subjects in the High-KR practice condition moved toward the target location (without a block) and, after the first five ofeverysix trials, received KR on a computer monitor; after every sixth trial no-kr was available. Subjects in the Fade-KR practice condition moved toward the target location and received KR presented with a schedule analogous to that for the Fade-Presentation condition. The retention phase consisted of an immediate retention test, 2 min after the end of the acquisition phase, and a delayed retention test, one day later. To assess the generalizability of learning, a transfer test was given immediately following the delayed retention test. For the transfer test, a different target location wasdisplayed on the computer monitor at the start of each 317
Winstein, Pah/, and Lewthwaite transfer trial. The retention and transfer tests each consisted of two 6-trial blocks without augmented feedback (i.e., no-block/no-kr). Apparatus andtask A standard aluminum angular positioning lever was anchored to a table (see Figure 1). Lever position was transduced by a linear potentiometer wired through an A-D board to a 386-25 MHz computer. A half-inch thick pegboard was fixed to the surface ofthe table and dowels were used to mark start and target positions. A plywood cover rested on the table to prevent subjects from viewing their arm movements, start, and target positions. The subject sat with the right forearm resting on the padded lever and grasped a vertical handle at the distal end. The arm was positioned comfortably with the shoulder and elbow in slight flexion, such that a position was where the lever and forearm were parallel to the subject's frontal plane. Custom software (Hary, 1991) was used to control the timing intervals, displays, and data collection. A color graphics monitor located 5 cm in front of the subject displayed the target location and KR when appropriate. The task required the subject to hold the lever and rapidly extend the forearm from eight different starting positions that ranged from to 45 to a target position located at 8 ; the transfer task target position was 1- Figure 1. Experimental setup. The outercirclesonthetable surface indicate the different startpositions. The one innercircle represents the peg hole for thetargetlocation where the block was placed for presentation trials.for subjects inthe knowledge of results (KRI conditions, thetargetlineand response linewere displayed afterthe response along with constant error(cei and thetrial number. For subjects in the presentation conditions, only thetrial number was displayed afterthe response. 318 cated at 5. Starting positions were randomized within each condition to assure that the task was to learn a target location independent of distance.' The movement time criterion required that the movement be completed within 8 ms. Subjects were verbally probed at four intervals during the acquisition phase regarding their specific thoughts prior to movement on the previous trial. Answers were recorded on audiotape for later analyses. Finally, subjects filled out two questionnaires regarding their thoughts about the learning of this motor skill; one was completed after the acquisition phase, the other after the transfer test. Procedure All subjects read and signed an informed consent. Subjects were seated facing the computer monitor with their right forearm positioned on the lever. Before the experimental session, each subjectwas familiarized with the experimental procedures and practiced moving the lever several times to a practice target different from that used in the experimental session. Specifics ofeach form of augmented feedback (physical block or KR) were explained according to the practice condition assignment. All subjects were informed at the start of the study that several retention sessions would be conducted in which they would have to move to the original position and a new target position without any augmented information. A trial was initiated with a 2-spremovement interval during which a graphic display of the target location appeared on the screen as a white radial line within a purple circle (see Figure 1). The center of the circle represented the axis of the lever arm. During this premovement interval, the experimenter indicated the feedback condition for the upcoming trial (i.e., block, no-block, error score, no-error score). Following this interval, the screen was cleared and two tones sounded sequentially, synchronized with a yellow warning light and a green start light that appeared on the screen. At the green start light, subjects began their movement to the target location. Reaction time was not measured; however, subjects were encouraged to move immediately after the green light. Ifthe movementwas not completedwithin 8 ms ofthe start signal, the trial was not valid and was repeated. Movement time was measured to determine if a trial was valid, but these data were not saved for later analyses. During the movement interval, the monitor remained blank. After the movement interval, for subjects in the presentation practice conditions, the monitor displayed the trial number in the upper right quadrant until the start ofthe next trial. The feedback display for subjects in the two KR conditions included a constant error (CE) score in degrees and a yellow radial line superimposedwith the white target line according to the specific feedback schedule ROES: December 1994
Winstein, Pohl, andlewthwaite (see Figure 1). Constant error was displayed on the monitor next to the trial number. The yellow line indicated the subject's end position and could be referenced to the white target position similar to the hands ofa clock. During a 3-s postmovement interval, subjects were instructed to remain at the end position. Then, with the subject's arm resting on the lever, the experimenter brought the lever to a new start position in preparation for the next trial. Start positions were randomized, and the intertrial intervals were held constant for all conditions. The beginning of the next trial was signaled by the appearance on the monitor of the premovement target display. Results For the acquismon phase, a 2 x 2 x 6 (Feedback Type x Frequency x Trial) ANOVA with repeated measures on the last factor was performed using absolute error (AE) as the dependent measure. This analysis used Trials 6, 12, 18, 24, 3, and 36 because these trials were performed under a common condition for all groups (no-block/no-kr). Using these common trials rather than a trial block analysis removed the artificial advantage of movement to a physical block by subjects in the presentation groups. For each retention phase (immediate and delayed) and the transfer phase, separate 2 x 2 x 2 (Feedback - 8 C) Q) --I- l- I- W High-Present 1 lj Fade-Present L. High-KR t>. Fade-KR 6 Q)-~ II) ~ -< 2 6 12 18 24 3 Acquisition Trials Figure 2. Group mean absolute error(ae) scores in degrees for acquisition trials under common feedback conditions (Le., no-kr/test trials). Note. Present =presentation; KR =knowledge of results. 36 Type x Frequency x Trial Block) ANOVAs with repeated measures on the last factor were performed using absolute constant error (ICEI) and variable error (VE) as the dependent measures. Absolute constant error is a measure of accuracy around the target location and, unlike AE, is independent ofve, which is a measure of within-subject variability around the subject's own blockmean. The two phasesofretention were analyzed separately in order to distinguish short-term and longer-term effects known to be important from previous work (e.g., see Schmidt, 1991). For all Ftests, significance was set at p ~.5, and the Greenhouse-Geisser degrees of freedom adjustment was used to compute the probability level for the repeated measure. Post-hoc linear comparisons using a Bonferroni correction were performed to determine the loci of significant main effects and interactions of interest. Effect size (ES) was calculated/ for significant between-subject factor effects. Acquisition Figure 2 shows AE group mean trial performance under common conditions. Subjects in all groups showed a reduction in AE with practice and attained a similar level of accuracy by the end of practice (i.e., Trial 36). The common trial analysis revealed a significant Trial effect, F(5, 18) =3.7, p«.1. There were no other main effects or interactions. The left side of Figure 3 (Block 1-6) shows that the ICEI Block means for the High-Presentation group are consistently low because five of every six trials had no-error. Therefore, 14 ~ 6- Cl Q) ~.. t:. g 1 W... 8 C a:l High-Present Fade-Present 12.. Hlgh-KR.z: Fade-KR / ~ ~ iii c 8 6- ~O Q) 4 \6-6V~ 'S ~ til 2 " ~~1~~.c «. "- ".-.-.-.-.-.,...---,,...---,,...---, 2 3 4 5 8 7 8 g 1 11 12 Acquisition Imm Del Transfer Figure 3. Group mean six-trial block absolute constant error(icei) scores in degrees for acquisition (Block 1-6), immediate retention (Block 7-8), delayed retention (Block 9-1), and transfer (Block 11 12). It should be noted thatthe Trial Block acquisition results are included for completeness only. Group comparisons across this phase were only performed onthetrial data due to the presence of thetargetblock. Note. Present = presentation; KR = knowledge of results; Imm =immediate; Del =delayed. ROES: December 1994 319
Winstein, Pohl, andlewthwaite group comparisons using Block means were not analyzed for these acquisition data. Retention The right side of Figure 3 shows the immediate (Block 7-8) and delayed (Block 9-1) no-feedback retention test ICEI means for the four practice conditions. Immediateretention test. The performance ofsubjects in the Presentation groups deteriorated over immediate retention blocks, while the performance of subjects in the KR groups was essentially unchanged. These results were reliable as indicated by the Feedback Type x Trial Block interaction, F(I, 36) = 6.47, p«.2, and followup tests showed a significant Block effect only for subjects in the Presentation groups, F (1, 18) = 6.98, P<.17, ES =.45. In addition, a Feedback Type x Frequency interaction was found, F (1,36) = 7.88, P<.9. Collapsed over the two immediate retention blocks, ICEI was lower for subjects in the High-KR group compared to thatofsubjects in the Fade-KR group; however, ICEI was lower for subjects in the Fade-Presentation group compared to that of subjects in the High-Presentation group. Post-hoc comparisons revealed that subjects in the High-KR condition had a lower ICEI than those in the Fade-KR condition, F (1,36) = 6.52, P<.2, but there was no difference between the performance ofsubjects in the High-Presentation and Fade-Presentation conditions. The VE findings showed a significant Block main effect as subjects' performance became more consistent over blocks, F (1,36) = 33.69, P<.1. There were no other main effects or interactions. Delayed retention test. The performance ofsubjects in the High-Presentation group was poorer than that of subjects in the other three groups whose performance was similar (see Figure 3, Block 9-1). A significant Feedback Type x Frequency interaction for ICEI was obtained, F (1,36) = 3.97, P=.5. Post-hoc analysis revealed that subjects in the High-Presentation condition (filled square symbol in Figure 3) performed significantly worse than those in the other three conditions, F(I,36) = 12.85, p «.2, ES=.95. No other main effects or interactions were found. A main effect for Block was obtained for VE, F(1, 36) = 18.17, p<.1. The performance of all subjects became more consistent across delayed retention. Transfer The right side of Figure 3 shows ICEI for the transfer test (i.e., Block 11-12). Analysis of subjects' performance in the transfer test revealed a significant main effect for Frequency, F (1,36) = 4.8, p «.4, ES=.49. Subjects who had practiced in the fade frequency conditions (open symbols in Figure 3) had a lower mean ICEI 32 than did subjects who had practiced in the high frequency conditions (filled symbols in Figure 3). Analysis of VE revealed main effects for Block, F(I,36) =5.92,p<.3,and FeedbackType,F(I, 36) =4.68, P<.4, ES =.48. Subjects in all conditions showed a decrease in variability over blocks. Subjects in the presentation practice conditions performed with less variability (M = 2.98 ) than those in the KR conditions (M= 3.7IO). Discussion The results support earlier motor learning findings using a presentation-test paradigm in which practice under conditions of frequent presentation trials was detrimental for learning compared with practice under conditions of frequent test trials (Hagman, 1983, Experiment 2). Ofparticular import, these results demonstrated that in the learning of a target location, there was an interaction between the type ofaugmented feedback and the frequency with which it was provided during practice (see Figure 3, Block 9-1). Frequent (83%) on-target experience during practice was detrimental for learning as evidenced by retention and transfer test performance. In contrast, a practice condition with the same frequency of less guiding feedback (High-KR) was not detrimental for learning as evidenced by retention test performance. In fact, practice in this high frequency, but less guiding, feedback condition was as effective for learning (as evidenced by retention test performance) as was practice in conditions with infrequent (33%) augmented feedback (on-target experience or KR). This suggests that the guiding properties of on-target information may interfere to a greater extent with information-processing operations important for retention than the guiding properties of KR when presented with the same high frequency. Further, these results suggest that the apparently detrimental effects ofheavily guided practice for learning can be offset if feedback is provided less frequently as in the Fade-Presentation practice condition. In most cases, practice conditions with high frequencies ofkr, relative to those with lower frequencies of KR, have been shown to be detrimental to learning, as evidenced by performance on delayed retention tests (Vander Linden, Cauraugh, & Greene, 1993; Weeks, Zelaznik, Thomson, & Williams, 1991; Winstein & Schmidt, 199, Experiments 2 & 3; Wulf, 1992; Wulf & Schmidt, 1989, Experiment 2). Here, practice in a condition with frequent KR (83.3%) was as effective for retention of a target location as a condition with less frequent KR (33.3%). However, the previous investigations demonstrating the detrimental effects of frequent KR have all used a 1% KR relative frequency schedule. It RaES: December 1994
Winstein, Pohl, andlewthwaite is possible that a KR schedule with slightly less than 1% frequency, such as that used here, is not as detrimental for retention as is one with 1% KR. In fact, subjects in the High-KR condition had six trials without augmented feedback during practice. These no-kr trials may have provided sufficient incentive and opportunity to overcome the guiding effects of the high frequency KR schedule. It should be noted that several previous KR studies have shown that practice with a 1% KR relative frequency schedule is as effective for learning, as indicated by performance on a delayed retention test, as one with less frequent KR (Schmidt & Shapiro, 1986, Experiment la; Sparrow & Summers, 1992, Experiment 1). In contrast to the retention test results, where there was an interaction between the frequency of augmented feedback and the level ofguidance provided by the feedback, the transfer test results using ICEI showed an effect of feedback frequency independent of the level of guidance. In the transfer test, both high relative frequency feedback groups performed poorly. These results support the view that processes that are optimal for retention may be different from those optimal for transfer (Lee, 1988; Morris, Bransford, & Franks, 1977). Bransford, Franks, Morris, and Stein (1979) stated that "acquisition-test relations that facilitate accuracy of retrieval are not necessarily equivalent to those that permit the types of refinements in understanding that facilitate subsequent transfer" (p, 35). In this light, a practice condition with a relatively high frequency of KR, while allowing processing appropriate for retention, did not appear to support processing appropriate for transfer to a new target location. It is interesting to note that the transfer test results for VE showed an effect of feedback type independent offrequency. Both KR feedback groups performed with greater variability. This suggests that augmented feedback that is more guiding during practice promotes more consistency in transfer performance than less guiding augmented feedback (KR). As such, the Fade Presentation condition promoted both accuracy and consistency during transfer. This finding was consistent with our predictions regarding the value of more precise on-target information provided less frequently. This study provides insight into the aspect of augmented information responsible for guidance-associated learning deficits. High-KR and High-Presentation practice schedules were similar in that they provided the same number of test trials (without the physical block or KR). They differed in the amount of guidance-kr provided arguably less guidance and potentially more difficultfeedback translation demands than a presentation trial to a physical block at the target. These two forms of augmented feedback during practice are thought to invoke different recall and recognition strategies for movement production. For example, RQES: December 1994 in practice conditions with a high frequency of KR trials, premovement retrieval operations must at least involve movement initiation and spatial endpoint planning (Meyer, Smith, & Wright, 1982). Responses from posttask questionnaires provided evidence for this; subjects in the High-KR group described attempts to replicate movements that had resulted in low error scores, concentrating on the feel of the arm position at the target location (Lewthwaite, Winstein, & Pohl, 1993). In contrast, in practice conditions with a high frequency of presentation trials, premovement retrieval operations may be minimal. The subject need only initiate the movement. Subjects in the High-Presentation condition described moving quickly to the physical block without having to concentrate on stopping at the target location. This occurred despite subjects being informed that a retention/transfer test without augmented feedback (KR, physical block) would be given. Practice conditions with a high frequency of presentation trials may promote the learning ofa different muscle activation pattern than thatrequiredfor the retention and transfer conditions when endpoint position must be generated from intrinsic sources. Previous work has shown that movements to a target location with and without a physical block are associated with different muscle activation patterns (Waters & Strick, 1981). Rapid movements to a physical block showed only an agonist muscle burst, while rapid movements to a target location without a blockwere associated with an agonist and an antagonist braking muscle burst (Waters & Strick, 1981). However, the potential existence of a similar acquisition-test condition effect in our experiment fails to explain the poor transfer performance of subjects in the High-KR condition. These subjects had ample opportunity to practice the agonist-antagonist muscle activation sequence required in transfer. This argues against the hypothesis that the different feedback practice conditions promoted the learning of different muscle activations. In general, the results provided support for the guidance hypothesis with respect to the schedule of augmented feedback opportunities. That is, practice in conditions with relatively infrequent exposures to augmented feedback was as good or better for learning than practice in conditions with relatively frequent exposures to augmented feedback. In addition, the effects on retention of the scheduling of augmented feedback appeared to depend on the extent ofguidance associated with each type of feedback. A relatively high frequency schedule of heavily guided practice was particularly detrimental for retention (ES=.95) compared to one with less frequent guided practice and ones with KR, regardless offrequency. Thus, the interaction ofthe type ofaugmented feedback and frequency appears to influence task-specific learning as revealed by retention performance. The frequency of augmented feedback 321
Winstein, Pohl, and Lewthwaite appears to have a moderate influence on generalizability (transfer) oflearning (ES=.49), as revealed by ICEI (Wulf & Schmidt, 1989), while the type of augmented feedback appears to have a moderate influence on consistency in transfer performance (ES =.48), as revealed by VE. Taken together, the results of this experiment suggest consideration ofthe combined contributions of the guiding properties of augmented feedback, the relative frequency with which it is provided, and the tests used to measure learning. References Annett.j., & Kay,H. (1957). Knowledge ofresults and "skilled performance." Occupational Psychology, 31, 69-79. Armstrong, T. R (197). Trainingfortheproduction of memorized movementpatterns (Tech. Rep. No. 26). Ann Arbor: University ofmichigan, Human Performance Center. Bjork, R A. (1988). Retrieval practice and the maintenance of knowledge. In M. M. Gruneberg, P. E. Morris, & R N. Sykes (Eds.), Practical aspects of memory, II (pp. 396-41). London:JohnWiley & Sons. Bransford, J. D., Franks, ]. J. Morris, C. D., & Stein, B. S. (1979). Some general constraints on learning and memory research. In L. S. Cermak & F.I. M. Craik (Eds.), Levels of processing in human memory (pp, 331-354). Hillsdale, NJ: Erlbaum. Hagman, J. D. (1983). Presentation- and test-trial effects on acquisition and retention of distance and location. journal ofexperimental Psychology: Learning, Memory, and Cognition, 9, 334-345. Hary, D. (1991). Template [Computer software]. Santa Monica, CA: Integrated Scientific Resources. Laabs, G.]. (1973). Retention characteristics of different reproduction cues in motor short-term memory. journal of Experimental Psychology, 1, 168-177. Landauer, T. K., & Bjork, R A. (1978). Optimum rehearsal patterns and name learning. In M. M. Gruneberg, P. E. Morris, & R N. Sykes (Eds.), Practical aspects of memory (pp. 625-632). New York: Academic Press. Lee, T. D. (1988). Transfer-appropriate processing: A framework for conceptualizing practice effects in motor learning. In O. G. Meijer & K. Roth (Eds.), Complex movement behaviour: The motor-action controversy (pp, 21-215). Amsterdam: Elsevier Science Publishers. Lee, T. D., White, M. A., & Carnahan, H. (199). On the role ofknowledge of results in motor learning: Exploring the guidance hypothesis. journal of Motor Behavior, 22, 191 28. Lewthwaite, R, Winstein, C.]., & Pohl, P. S. (1993). [Questionnaire responses in motor skill acquisition]. Unpublished raw data. Magill, R A. (1993). Augmented feedback in skill acquisition. In R N. Singer, M. Murphey, & L. K. Tennant (Eds.), Handbook of research on sportpsychology (pp.193-212). New York: Macmillan. Meyer, D. E., Smith,J. E. K., &Wright, C. E. (1982). Models for the speed and accuracy of aimed movements. PsychologicalRevif!W, 89, 449-482. Morris, C. D., Bransford,J. D., & Franks.j.]. (1977). Levels of processingversus transfer appropriate processing.journal of Verbal Learning and Verbal Behavior, 16, 519-533. Newell, K. M., Morris, L. R, & Scully, D. M. (1985). Augmented information and the acquisition of skill in physical activity. In R L. TeIjung (Ed.), Exercise and sportsciences reoieus (Vol. 13, pp. 235-261). New York: Macmillan. Salmoni, A. W., Schmidt, R A., & Walter, C. B. (1984). Knowledge of results and motor learning: A review and critical reappraisal. Psychological Bulletin, 95, 355-386. Schmidt, R A. (1991). Frequent augmented feedback can degrade learning: Evidence and interpretations. In G. E. Stelmach &J. Requin (Eds.), Tutorials in motorneuroscience (pp. 59-75). Dordrecht, The Netherlands: Kluwer Academic Publishers. Schmidt, R A., & Shapiro, D. C. (1986). Optimizingfeedbackutilization in motorskill training (Report No. 1/86). Alexandria, VA; U.S. Army Research Institute (Basic Research). Schmidt, R A., Young, D. E., Swinnen, S., & Shapiro, D. C. (1989). Summary knowledge of results for skill acquisition: Support for the guidance hypothesis. journal of Experimental Psychology: Learning, Memory, and Cognition, 15, 352-359. Sparrow, W. A., & Summers.j.]. (1992). Performance on trials without knowledge ofresults (KR) in reduced relative frequency presentations of KR. journal of MotorBehavior, 24,197-29. Swinnen, S. P., Schmidt, R A., Nicholson, D. E., & Shapiro, D. C. (199). Information feedback for skill acquisition: Instantaneous knowledge of results degrades learning.journal of Experimental Psychology: Learning, Memory, and Cognition, 16,76-716. Vander Linden, D. W., Cauraugh, J. H., & Greene, T. A. (1993). The effect of frequency of kinetic feedback on learning an isometric force production task in nondisabled subjects. Physical Therapy, 73, 79-87. Waters, P., & Strick, P. L. (1981). Influence of "strategy" on muscle activity during ballistic movements. Brain Research, 27, 189-194. Weeks, D. J., Zelaznik, H. N., Thomson, B., & Williams, K. (1991). Disproving the disprooffor the guidance hypothesis. [Abstract]. Proceedings of the North AmericanSociety for the Psychology of Sport and Physical Activity (p, 127). Monterey, CA: NASPSPA. Winstein, C.]., & Schmidt, R A. (199). Reduced frequency of knowledge of results enhances motor skill learning. journal of Experimental Psychology: Learning, Memory, and Cognition, 16,677-691. Wulf, G. (1992). The learning of generalized motor programs and motor schemata: Effects ofkr relative frequency and contextual interference. journal of Human Movement Studies, 23, 53-76. Wulf, G., & Schmidt, R A. (1989). The learning of generalized motor programs: Reducing the relative frequency of knowledge of results enhances memory. journal of Experimental Psychology: Learning, Memory, and Cognition, 15, 748-757.. 322 RQES: December 1994
Winstein, Pohl, andlewthwaite Notes I. The choice to dissociate target location from distance moved was based on earlier work (Laabs, 1973) in which presentation trials were found to promote better retention of location than distance. 2. Effect sizes were calculated using the following equation: (MI - M2)/SD, where MI and M2 represent group means (e.g., High-Presentation, Fade-Presentation) and SD is the square root of the mean square error between-subjects from the highest order ANOVA Authors' Notes The authors thank Dan Corcos, Tim Lee, Ben Sidaway, and Maureen Weiss for helpful comments on an earlier version ofthis paper. Address correspondence to Carolee J. Winstein, PhD, Department of Biokinesiology and Physical Therapy, 154 East Alcazar St., CHP-155, University ofsouthern California, Los Angeles, CA 933. E-mail: winstein@hsc.usc.edu RDES: December 1994 323