Hodges and Franks

Modelling coaching practice the role of instruction anddemonstration

NICOLA J HODGES1 and IAN M FRANKS21Research Institute for Sport and Exercise Sciences Liverpool John Moores University 15plusmn 21 Webster Street

Liverpool L3 2ET UK and 2School of Human Kinetics University of British Columbia 210 War Memorial Gym6081 University Boulevard British Columbia V6T 1Z1 Canada

Accepted 16 February 2002

In this paper we review the empirical literature pertaining to the ethorn ectiveness of instructions and movementdemonstrations Initially we examine existing theories and approaches that try to explain the process of skillacquisition so as to determine implications of these theories for instructional provision This is followed by anevaluation of studies in the motor learning literature in which pre-practice information has been manipulatedExplicit learning strategies are contrasted to implicit and discovery learning methods and current explanationsfor instructional ethorn ects are discussed in terms of such mechanisms as ethorn ects-related attentional focus andmovement variability In the reg nal sections we review data from our own laboratory where pre-practiceinformation has been manipulated during the learning of a novel bimanual coordination task From thesestudies proposals are made to try and explain how pre-practice information works to ethorn ect the process ofskill acquisition including the selection and execution of a response and the processing of associated feedbackAn important role is given to the existing skills of the learner in understanding the instructions and performingthe desired movement Finally we suggest some practical implications of this empirical evidence for the teachingof motor skills

Keywords augmented information motor learning skill acquisition

Introduction

Performance analysis of sport has been used primarilyto inform the coaching process Objective informationabout an athletersquo s performance is used by the coach todesign the practice environment and subsequentlyaid in the modireg cation of athlete behaviour Thereforethe `practice sessionrsquo itself can be considered a criticalelement in the development of skilled athletic perform-ance During this `practice sessionrsquo ethorn ective instructionmay be crucial to the pursuit of optimal sporting per-formance as the more ethorn ective the instruction themore the instructorrsquos role will benereg t athlete perform-ance Such instruction requires the application of skillsthat range from the planning and organization of learn-ing experiences to the presentation of instructional andfeedback information Hence one of the dereg ning rolesof the coach is that of instructor As an instructor thecoach is responsible for teaching the athlete what to do

Author to whom all correspondence should be addressed e-mailifranksinterchangeubcca

how to do it and hopefully how to do it well This mayinvolve providing instruction about optimal movementpatterns or feedback on errors relating to specireg c taskgoals Instruction can be provided in various modalitiesand at particular times For example the coach mayprovide a visual demonstration to the learner before anyexperience with the to-be-learned skill provide instruc-tion verbally during or after execution possibly relatingto performance of the whole skill or home in on specireg caspects of the skill Here we investigate the role that pre-practice information plays in learning a motor skill Weoperationally dereg ne instruction and demonstration asinformation given to athletes by the coach at the start ofpractice (ie before physical practice begins) or duringpractice but independently of performance Informationprovided during or after performance is consideredfeedback and will not be dealt with in detail here

Although there is considerable empirical evidenceof the correct methods for coaches and teachers inproviding feedback to athletes and learners generally(More and Franks 1996 Franks 1997 Hughes andFranks 1997) less is known about the ethorn ectiveness of

Journal of Sports Sciences 2002 20 793plusmn 811

Journal of Sports Sciences ISSN 0264-0414 printISSN 1466-447X online Oacute 2002 Taylor amp Francis Ltdhttpwwwtandfcoukjournals

instruction and demonstration Despite the shortage ofunequivocal empirical evidence to support the provisionof pre-practice information most coaching educationprogrammes sanction its use and consider it to be a keyfactor in teaching motor skills A physical education(PE) researcher specialist commented Ìt would be astrange PE lesson indeed it would be a rare PE lessonif some form of physical demonstration was not shownto some or all of the pupilsrsquo (Blake 1998 59)

In fact little scientireg c consensus exists about the rolepre-practice information plays in the learning processbeyond the sometimes necessary role of relaying to thelearner the goals of the task This gap in knowledge isparticularly interesting given that augmented feedbackduring and after skill execution has received markedattention in the motor learning literature In their mostrecent motor control and learning textbook Schmidtand Lee (1998) echo these sentiments by concludingtheir section on pre-practice conditions with thefollowing statement few investigations have beencarried out on the nature of instructions and muchwork remains to be done in assessing their value inlearningrsquo (p 289)

In this paper we review research that has focusedon the ethorn ectiveness of pre-practice information forlearning motor skills and suggest some guidelines for`best practicersquo coaching behaviour that is based on thisempirical evidence First we discuss the theory under-lying these research reg ndings and the implications ofthese theories for providing pre-practice informationWe include a discussion of traditional motor learningtheories (eg Adams 1971 Schmidt 1976) and otherapproaches that posit a primary role for movementrepresentations (eg Fitts and Posner 1967 Bandura1986) Additionally we discuss less prescriptiveapproaches with an emphasis on the constraints in themotor system that are responsible for the emergingmovement

We then review empirical reg ndings and contrastexplicit learning strategies with more implicit discoverylearning environments Dithorn erent performance goals areexamined such as criterion templates demonstrationsand verbal instructions and we try to determine themechanisms underlying these various instructionalmethods including focus of attention and movementvariability These various performance goals how theyinmacr uence the emerging movement and in turn areinmacr uenced by pre-existing skills and the results of themovement are subsequently illustrated in a macr ow dia-gram (see Fig 1) In the reg nal sections we present aseries of experiments from our laboratory in which weexamined the ethorn ects of dithorn erent instructional variableson the learning of novel bi-manual coordinationmovements We discuss how various performancegoals athorn ect the selection of a movement response and

the subsequent processing of feedback We end with atable that outlines some of the practical implications ofthe research reviewed for the teaching of motor skills

Motor learning theory and implications forinstruction and demonstration

Classic theories of skill acquisition

The question of how people acquire novel motor skillshas long been a topic of interest This dates back to theearly work of the behaviourists (eg Thorndike 1927Skinner 1953) when the outcome of the movementand the determination of environmental contingencieswere of primary signireg cance With the cognitiverevolution came an associated interest in the processesresponsible for learning new behaviours and conceptsrelated to information processing In motor behaviourtwo theories have received most attention These areAdamsrsquo (1971) closed-loop theory of motor controlhighly inmacr uenced by the reg eld of cybernetics andSchmidtrsquos (1975 1976 1988) schema theory wherebymotor programs are proposed to be responsible formovement execution much like software programs incomputers

Schmidt and Adams proposed that learning results inthe development of two representations of the move-ment in memory either actual memory traces (Adams1971) or abstractions called schemasrsquo (Schmidt 1975)These representations are responsible for initiating themovement and evaluating its ethorn ectiveness both duringand after execution The perceptual trace or recognitionschema is responsible for evaluating the ethorn ectivenessof the response and therefore acts as a referencemechanism to compare the actual sensory con-sequences with those desired After practice the per-ceptual trace represents those sensory or feedbackcharacteristics of the correct response whereas therecognition schema allows the learner to estimate theexpected sensory consequences of an action becausegeneral (invariant) features of the movement are storedGiven the importance of the detection and recognitionof errors in both Adamsrsquo and Schmidtrsquos theoriespre-practice information which is designed to rereg nethe movement template (trace or schema) should beimportant in aiding this process Such information thatallows the learner to extract easily and eyacute ciently theexpected response either in verbal or visual form wouldarguably aid the development of a reference with whichfeedback about the movement can be compared (seealso Swinnen 1996) This assumption only holdshowever if a movement template is there to begin

Pre-practice information might easily inmacr uence theresponse that is evoked or selected in the early stages of

794 Hodges and Franks

skill acquisition when the learner is unable to performthe desired action If the learner has an existing generalprogram for the movement as a result of previousexperience with similar skills then information relatingto correct parameter selection would possibly beimportant (eg how much force to apply to the move-ment to achieve a required distance) Pre-practicedemonstrations might also provide elaboration for amovement reference after some practice at the requiredskill has been attained If the general motor program hasnot yet been established but the response componentsare relatively simple then perhaps information con-cerning invariant features should be emphasized ormade salient (eg a demonstration showing the correctorder of components) However when the movementresponse has not been performed before or is suyacute ci-ently novel it is diyacute cult to know what type of informa-tion would be most ethorn ective in aiding the developmentof a new motor program No explicit mention was madeby Schmidt or Adams about how these movement repre-sentations become established Given that response-produced sensory feedback has been central to thedevelopment of these models and that this informationforms the basis of a movement store it is unclear howsensory information in the absence of action (ie ademonstration) would or could be used (for a discussionof these limitations see Schmidt 2001)

Other representational-based accounts of skill acquisition

Variations to representational-based theories andmechanisms have been proposed to underlie skill acqui-sition Gentile (1972) proposed that acquisition of ageneral motor pattern or what she referred to as gettingan idea of the movementrsquo preceded either rereg nement ordiversireg cation of the general pattern depending on theenvironmental conditions Whiting and den Brinker(1982) also distinguished two motor images that areacquired through practice First the image of the actrsquo where the learner is primarily concerned with attainingthe task goal and acquiring the form of the movementsecondly a model of the forces to be overcome called`the image of achievementrsquo Franks and Wilberg (1982)found that early learners of a tracking task developed ageneral image of the required movement plusmn as evidencedby drawings plusmn much like the image of the actrsquo proposedby Whiting and den Brinker

During this early stage of motor skill acquisitionunderstanding the task is the primary goal Fitts andPosner (1967) claimed that during this so-called cogni-tive stage instructions and movement demonstrationsare most important in aiding the development of anexecutive program This cognitive stage which isbelieved to be primarily verbal shares similarities withstage models of acquisition proposed by Anderson

(1982) and Adams (1971) whereby procedures andrules for performing a task get proceduralized orbecome more automatic and less verbal with practiceIn the early stage of acquisition Fitts and Posner (1967)proposed that processing of feedback and attending toimportant cues is critical for learning Thereforeinstructions that serve to emphasize these cues arerecommended such as attending to the feet whendancing Although very inmacr uential in motor learningFitts and Posnerrsquos theory has been somewhat limitedto observations of individuals learning information-processing skills with primarily cognitive or perceptualdemands such as language skills or reaction timestudies rather than diyacute cult motor componentsAdditionally much of the theorizing was based onanecdotes and interviews with instructors Indeed somerecent research has questioned these earlier speculationsas to good practice methods For example Wulf et al(1998) found that during the acquisition of novelwhole-body movements plusmn ski-like movements on asimulator plusmn encouraging participants to pay attention totheir limbs was detrimental to the acquisition process

Fitts and Posner (1967) also attributed a primary roleto habits in inmacr uencing the learning of new skills andproposed that `The learning of skills is largely a matterof transfer of prior habits to new situationsrsquo (p 19) Sub-sequently instructions that serve to elicit previouslylearned behaviours were proposed to be important inactivating the appropriate `cognitive sets and expec-tanciesrsquo (p 20) and providing a language for under-standing the new task However in which circumstancesthis instruction would be appropriate and benereg tacquisition was not elaborated although they arguedthat `These general sets will contain aspects bothappropriate to the new learning and inappropriate to itrsquo(p 20)

Another inmacr uential motor skills researcher wasEdwin Fleishman In several studies primarily directedat aiding training and selection protocols for the military(eg Fleishman 1957 Parker and Fleishman 19591961) Fleishman and colleagues identireg ed changingpatterns of ability as a function of practice For exampleParker and Fleishman (1959) found that spatialplusmn verbalability as assessed by pencil-and-paper tests was animportant predictor of performance early in acquisitionof a motor skill supporting the work of Fitts andPosner As practice progressed this cognitive ability wasreplaced by psychomotor ability reg nally coordinationskill accounted for the most variance at the end ofpractice on a tracking task

Parker and Fleishman (1961) subsequently went onto manipulate verbal instruction relevant to these basicabilities on a tracking task Their aim was to devise task-appropriate training strategies in which basic abilitieswere maximized at dithorn erent stages throughout practice

Instruction and demonstration 795

Performance based on this training strategy was foundto be better than that of a non-instructed controlgroup although the dithorn erences were not so markedwhen comparisons were made with a `common-sensersquotraining group Based on these reg ndings it is temptingto conclude that instructions about coordination skillsshould perhaps be given later in practice and demandson cognitive resources kept to a minimum early inacquisition Decreasing cognitive demands early inacquisition would enable problem-solving capabilitiesto be devoted to the task rather than implementingand understanding instructions and movementdemonstrations

More recently Ackerman and colleagues (egAckerman 1988 1992 Ackerman and Cianciolo2000) expanded on Fleishmanrsquos original ideas and pro-vided an integrated theory of individual dithorn erencesbased on the skill acquisition framework of Fitts andPosner (1967) and Shithorn rin and Schneider (1977)Shithorn rin and Schneider (1977) proposed that shifts inprocessing demands from controlled or highly cognitiveto more automatic as a function of practice only arosewhen the stimulusplusmn response components of the taskwere consistent Using a complex air-trayacute c control taskwhich integrated both simple and complex task com-ponents Ackerman (1992) showed that general cogni-tive abilities were most predictive of performance at theearly stages of skill acquisition for task components withconsistent information-processing demands Howeverfor more complex task components cognitive abilitiesmaintained high performance predictability throughoutpractice The implications of this work is that individualdithorn erences in general cognitive functioning will playa signireg cant role throughout the acquisition processand that for more complex tasks or task componentsthe ability of the learner to translate and use informationethorn ectively will be important

Another highly inmacr uential theory of learning againbased on the development of movement representa-tions is the social learning theory of Bandura (19691977) This cognitive-behavioural theory of learninghas been the predominant theory guiding researchinto learning through observation plusmn so-called modellingA spatial or imaginal system and a verbal system wereproposed to underlie the development of movementrepresentations Verbal mediation allowed for a repre-sentation of the modelrsquos behaviour and enabled cate-gorization The spatial representation was proposed tobe most important early in the development stage whenthe capacity to verbalize was not available or not welldeveloped Both visual and verbal instructions anddemonstrations were believed to inmacr uence learning insimilar ways in cognitively developed adults Bandura(1986) proposed several cognitive processes underlyingethorn ective modelling The learner must selectively attend

to the relevant information retain this information havethe capability for using the information and have thedesire or motivation to imitate the behaviour

In terms of pre-practice information these proposalshave important implications for both attentional focusspecireg cally the determination of `relevantrsquo informationand movement capability Obviously there must be noapparent physical limitations to modelling the desiredmovement although whether this implies that themovement response should already be an existing partof an individualrsquos movement repertoire is questionableIf a person cannot perform the response on the reg rstpractice trial does this mean they will not have thecapability to use the information Perhaps one of thecriticisms of Bandurarsquos work is the simplistic nature ofthe response that has been modelled Often a response isrequired that places more demands on the cognitiveaspects such as remembering a specireg c sequence ratherthan learning to perform complex motor components(eg Carroll and Bandura 1982 1985 1987) Thisresearch will be discussed in more detail in the nextsection

Alternative non-representational accounts of learning andperception

Although motor learning research has been predomin-ately explained by the development and strengtheningof movement representations that guide productionthere are alternative theories Dynamical systemstheory (see Kelso 1995 for a review) seeks to explainbehaviour by variables that bring about a change inmovement as such movement is seen as a consequenceor an emergent property of these variables In thisway movement solutions are discovered as a result ofexternal variables or constraints such as feedback fromthe environment and demands on speed as well asinternal constraints such as prior learning and thephysical structure of the motor system Newell (1986)emphasized the importance of the task in constrainingand shaping the reg nal movement Task requirementswhich may be specireg ed by instructions and demonstra-tions act to determine the movement but only againstthe backdrop of the internal and external constraintsthat were previously detailed

Given the emergent nature of behaviour a reducedemphasis on movement form as the instructed goal ofthe action might be expected (see Davids et al 1994)Less direct methods of teaching such as changing thephysical conditions in the environment would helpbring about change in movement form Although boththe dynamical approach and more representational-based accounts of learning have their own strengths andweaknesses dynamical systems theory othorn ers new toolsto tackle the issues of what is learned and what methods


best facilitate this learning process Through the studyof inter-limb coordination Kelso and colleagues (egKelso 1984 Haken et al 1985 Kelso et al 1986Schrdquoner et al 1992) have identireg ed stable modes ofcoordination that allow the learning of novel patternsto be examined in comparison to preferences for pre-existing stable movements Knowledge of these existingbehaviours othorn ers a way of understanding how pre-practice information inmacr uences the selection of an initialresponse and the resulting rereg nement and change inthese initial behaviours

The movement pattern is described or quantireg ed byan order parameter or collective variable In inter-limbcoordination this order parameter has been the relativeposition of one limb in relation to the position ofthe other limb at predetermined temporal and spatialmarkers Two basic patterns dominate bi-manual co-ordination and are produced in a reliable and stablefashion These are symmetrical in-phase movementscharacterized by simultaneous macr exion and extensionof homologous muscle groups (0deg relative phase)and asymmetrical anti-phase movements which aredescribed by alternating macr exion and extension ofhomologous muscles (180deg relative phase)

In several experiments the stability of these patternshas been demonstrated through manipulation of speedconstraints As movement frequency is increased theanti-phase pattern becomes less stable than the in-phasepattern This resulting variability in the observedmovement can lead to a transition from or reorganiza-tion of the anti-phase pattern to in-phase (Kelso1984) Given that two people who sat facing each otheralso showed similar transitions between limbs (Schmidtet al 1990) the implication is that environmentalconstraints alone (ie vision) are suyacute cient to elicit thischange in behaviour The instructions given to partici-pants have been shown to play a role in inmacr uencing thetime and indeed exhibition of these transitions Instruc-tions to a performer to resist intentionally switching toanother movement pattern act as higher-order taskconstraints inmacr uencing the observed movement Lee etal (1996) showed that under instructions to resist aswitch from anti-phase to in-phase coordination at highmovement speeds participants could indeed avoidany transitions to the in-phase pattern The diyacute cultyin maintaining the anti-phase pattern however wasindexed by high variability in the resulting movementThis reg nding illustrates the importance of takinginto consideration the task demands environmentaldemands and the existing skills of the individual whenexplaining and predicting behaviour

One of the important concepts of dynamical systemstheory and associated research into coordinationdynamics is the notion of stability and its complementvariability Dynamic systems need to have the capacity

to change and therefore some variability in a responseis necessary to promote change Bernstein (1967)proposed that the process of skill acquisition is oftenaccompanied by increased variability in the constituentlimbs and joints underlying the movement and thatto discover new patterns of coordination stability has tobe sacrireg ced Previously stable conreg gurations of jointsare freed to allow a more skilful performance (seeArutyunyan et al 1968 Vereijken et al 1992b whohave demonstrated this progression in sport skills) Animplication for early instruction is that it is importantto encourage initial variability in the movementresponse or at least not to discourage it such thatstable adherence to non-desirable movement patterns isavoided

Given the importance of the environment and inparticular visual perception in constraining or dictatingthe movement there has been a move to examine theinmacr uence of demonstrations on motor learning from aconstraints-led perspective This approach has beenfounded on Gibsonian principles of direct perception(eg Gibson 1979) where perception and action arebelieved to be integrally coupled As such the performeris able to pick-up directly information of functionalrelevance to the task at hand This direct pick-up ofinformation is not believed to be mediated by cognitivemechanisms which enable the learner to translate theobservation into action (cf Bandura 1986) It has beenproposed that certain environmental invariances specifyaction for the performer and that with learning theperformer becomes better attuned to this informationrather than particular details of the movement There-fore an important component of learning is beingable to use adequately this information to reg ndrsquo and`guidersquo optimal solutions to motor problems (Fowlerand Turvey 1978) An implication for pre-practicebehaviour is that the content of a movement demonstra-tion becomes extremely important There is a greateremphasis on what information is picked up from observ-ing a skilled model rather than how the informationshould be presented or how it is used (see Scully andNewell 1985) To date however empirical researchhas shown limited support for the idea that onlyinvariant features of the movement such as movementkinematics portrayed only as point lights are helpfulfor learning (see McCullagh and Weiss 2000) Giventhe strong coupling between perception and actioncritical perceptual invariants which arguably specify therequired action may not become available to perform-ers until a certain mastery has been attained Both theaction constrains the perception and the perceptionconstrains the action Teaching individuals thereforeto recognize critical information may prove diyacute cultGoulet et al (1989) remarked upon perceptual dif-ferences in tennis as a result of experience and rejected


the idea that these skills can be taught to less skilledplayers Additionally as will be remarked on later thereis evidence that the detection of perceptual invariance istypically acquired implicitly without conscious aware-ness of regularities on the part of the learner (egLewicki et al 1992)

Empirical evidence from research into theethorn ects of pre-practice information on skillacquisition

Criterion templates and optimal movement patterns

The importance of knowing the task goal was demon-strated in a simple study by Newell et al (1990)who required their participants to produce familiaror unfamiliar drawing shapes When the task require-ment was a familiar pattern (a circle) the participantsneither required this goal-related information to pro-duce the correct pattern nor required detailed per-formance feedback about how the movement wasperformed plusmn what has been termed knowledge of per-formance Rather knowledge about the success oroutcome of the movement typically referred to asknowledge of results was suyacute cient However if thetask goal was unfamiliar this criterion informationwas required during practice to produce benereg ts inacquisition and in a no-feedback retention test Newellet al concluded that without knowledge of what shouldbe produced (the task goal) no reference would beavailable to compare movement-related feedbacknegating any benereg ts of such feedback

Motor skills have often been distinguished by thesimilarity between the outcome goal and the movementrequired to attain the goal Two dithorn erent classes ofmovement have been proposed based on this distinc-tion so-called open and closed skills although they arecommonly viewed to lie on a continuum (see Gentile1972) For closed skills the goal of the task and how thetask goal is to be achieved are the same For example indiving achieving a dive in a specireg ed way is the goal ofthe action Open skills in contrast commonly haveconmacr icting goals where a solution to a particular taskgoal may be attained in many ways For example apartfrom the rules of the game that dictate which parts of thebody can be used to score a goal there are no con-straints placed on the way a goal is scored in soccernetball or hockey Although these distinctions existGentile (1972) proposed that the initial stage of skillacquisition is probably similar across open and closedskills In both cases the learner is required to reg nd a wayof organizing their motor system to produce a desiredoutcome and determine the appropriate information forrealizing this Additionally in many open-type skills thecoach or experimenter may turn the action into a closed

skill by making the achievement of a particular move-ment variable plusmn such as form or movement amplitude plusmnthe goal of the task In this instance optimal movementtemplates or criterion movement patterns will becomean additional goal for the learner and may take prece-dence at the expense of task attainment plusmn this hasbeen referred to as goal confusion (see Gentile 1972Whiting and den Brinker 1982) For example correctpositioning of the feet and upper body may be attainedduring a corner kick but this may be at the detriment ofachieving the required height and distance on the ball

One of the ethorn ects of trying to copy movement tem-plates has been an increase in the consistency of themovement (eg Newell et al 1990) Although thisconsistency may be desirable for closed skills whereconsistent production of the same movement patternis required in environments where the task demandsare continually changing this consistency may not bedesirable (Gentile 1972) Given the vast array of pos-sible sensory experiences and changing environmentalconditions that may be encountered by a tennis playerfor example varied practice conditions will be necessaryfor ethorn ective performance during match-play Perhapsmore importantly task consistency is only desirable ifthe skill is performed correctly Early in practice whenthe individual lacks the skills to perform the movementand is possibly biased towards other more easily attain-able movements (eg previously developed habits frompractice in other movement skills) this consistency islikely to hinder acquisition Lee et al (1995) examinedthe learning of a new bi-manual coordination patternand found that early in acquisition initial variabilityin the attempted response preceded acquisition of anew coordination pattern Because of pre-existingstable preferences to undesirable movement patternsvariability in the attempted movement was needed tofacilitate the break from these patterns Lee et al con-cluded that early in practice high variability might bedesirable indicating exploration of new coordinationpatterns and the break from old ones

Whether optimal movement patterns exist and there-fore should be used to teach new motor skills is still amatter of contention in many sports Newell et al (1985238) claimed that Às the optimal co-ordination func-tion for most tasks has never been determined formallyit typically remacr ects the intuitions of the advisor aboutwhat pattern of co-ordination is optimal andor mimicsthe co-ordination pattern that other successful perfor-mers have utilisedrsquo Given the vast dithorn erences betweenindividuals in previous practice experiences physicalstature or reg tness it should not be too surprising thatsimilar performance standards are not achieved insimilar ways What may be optimal for one person witha certain skill may not be optimal for another Evenin very simple skills such as hammering a nail


considerable intra-individual variability in the move-ments of the joints compared with the end-point of themovement plusmn the tip of the hammer-head plusmn has beennoted even after many years of performing the skill (seeLatash 1996)

In a recent experiment in which participants wererequired to coincide a movement reversal of the armwith a dynamic stimulus display plusmn somewhat akin tohitting an arriving ball plusmn Brisson and Alain (1996) foundno relationship between outcome and kinematicmarkers (ie the way the movement was performed)Furthermore in a retraining experiment it made nodithorn erence to performance scores which were calculatedbased on speed and accuracy whether participantspractised with a best participantrsquo s template or their ownmovement template In fact to achieve consistency inperformance an individually based movement templatewas more benereg cial than an optimally determinedone These reg ndings serve to question the underlyingethorn ectiveness of optimal movement templates in teach-ing motor skills and at the very least caution againsttheir automatic use when teaching new motor skills

It has also been argued that expert templates ormovement demonstrations cause people to try to repli-cate the solution without trying to reg nd their ownsolution Someone elsersquos solution may not be wellunderstood by learners and they have no supportingknowledge base to make it useful Higgins (1991) callsthis training rather than learning Bernstein (1996) hasalso remarked on the importance of repetition withoutrepetition plusmn that is the repeating of the means of solvinga problem rather than repeating the solution

In the preceding examples the main instructionaltechnique has been to provide kinematic movementtemplates that prescribe a certain way for performing askill In the next section visual demonstrations are dis-cussed in reference to their ethorn ects on the acquisition ofsimple motor skills

Demonstrations and modelling

The empirical evidence supporting the ethorn ectivenessof demonstrations as a teaching tool in motor skilllearning has been mixed (Newell 1981) The results ofthe research obviously depend somewhat on the typesof tasks that have been examined and subsequentlythe type of learning required When the task is rathersimple or the learner has some prior knowledge of thecriterion information to be conveyed then demonstra-tions are not useful In the words of Newell (1981 537)`When the information to be conveyed through demon-stration is redundant there is by dereg nition no informa-tion transmitted to the learnerrsquo Whether information isredundant will depend on both the task and the skills ofthe learner If the task is to make a familiar shape or to

perform a common task such as clapping it is unlikelythat a demonstration or a reference template will pro-vide any new information to the learner If a certainmethod of clapping is required then it is likely thatadditional visual and verbal information will be helpfulor indeed necessary for performance and learning

Despite the cases when the task is relatively simpleor the individual has some prior knowledge of the taskgoal movement demonstrations often play a criticalrole in specifying the requirements of the goal to theindividual The importance of this pre-practice infor-mation has received most support when tasks have beenexamined that require the `stringing togetherrsquo of move-ments that are already part of a personrsquos movementrepertoire Remembering a specireg c sequence of move-ments when learning a gymnastic routine for exampleis very dithorn erent to learning a specireg c gymnastics skillsuch as a cartwheel In the laboratory a common motorlearning task has been one of learning a sequence ofkey-presses on a keyboard For example Howard et al(1992) examined whether observing dithorn erent patternsof key sequences would enhance serial pattern learningcompared with a group who responded with actual key-presses during practice They found no dithorn erencesin overt responding between the groups as measuredby reaction time and prediction of the next key in asequence In these tasks when the learner can alreadyperform the motor skills required of the task but has tolearn the correct order in which to produce them thedemands on learning are predominantly cognitive ratherthan motoric The learner does not require experienceof these simple movements only cognitive elaborationof the to-be-remembered sequence

As detailed earlier Bandura has been inmacr uentialin directing research into the processes underlyingobservational learning A key feature of his theory isthat of attention Attention to the modelled action isnecessary for the development of a representation of themovement skill in memory either visually or verballywhich will serve as a reference of correctness to guidethe action when required However learning modelshave been shown to be just as good as expert modelswhich demonstrate correctrsquo performance and in somecases more benereg cial for learning (eg Lee and White1990 McCullagh and Caird 1990 Lee et al 1991)These reg ndings lead to the suggestion that somethingother than how to perform the action correctly is beingpicked up by observers of motor skills When watching alearning model learners are actively involved in theproblem-solving process and therefore are concernedwith goal attainment not just copying the model Theyreceive variable information across trials rather thanthe repetition of one possible solution Subsequentlywhen watching a learning model observers are notmerely trying to copy performance but rather reg nding


out what does and does not work Indeed mentalpractice plusmn another form of covert rehearsal plusmn has beenshown to have similar learning benereg ts as observationalpractice (see Feltz and Landers 1983 for a review)suggesting that cognitive processes are mediating thesebenereg ts

Movement demonstrations have been shown to beuseful in a variety of tasks that arguably have a highcognitive component (eg Carroll and Bandura 19821985 1987 Doody et al 1985 Ross et al 1985) Instudies of the ethorn ectiveness of movement demonstrationon learning the correct sequence and spatial require-ments of arm movements Carroll and Bandura (19821985 1987) found that demonstrations served toenhance the memorial representation of the requiredsequence This benereg t was shown in both measures ofreproduction and recognition accuracy Anothersomewhat dithorn erent line of evidence for this conclusioncomes from work examining cognitive strategies under-lying expert performance Figure skaters and balletdancers (Deakin 1987 Starkes et al 1987) show verygood recall for structured dance sequences acquired bywatching a model performing the sequence Experi-enced performers typically use hand movements duringobservation to `markrsquo the steps which serve to act as amarker for the whole-body position later in recall Thestrategy of marking the steps and therefore memory forthe correct sequence is enhanced by a demonstrationnot the acquisition of the individual moves that makeup the sequence although this was not examined byDeakin and co-workers

Observational learning studies have typically notrequired the learner to acquire a new movement pat-tern but rather rereg ne an existing movement Vogt(1996) has referred to the reproduction of tasks alreadywithin a personrsquos behavioural repertoire as imitationlearningrsquo in contrast to òbservational learningrsquo whichis the learning of newly acquired skills Despite thisdereg nition observational learning studies have typicallynot involved learning new skills Often learning has beena case of rereg ning old ones by scaling to a new taskparameter

It has also been shown that movement demonstra-tions can aid the learner in adopting strategies that areassociated with skilled performance (eg Martens et al1976 Kohl and Shea 1992) Kohl and Shea (1992)found that merely observing a group perform a trackingtask led to the development of a control strategy similarto that of a physical practice group That is theseindividuals showed a strategy characterized by rapidresponding not waiting for response-produced feed-back to initiate the next response However the accur-acy of responding was poorer than that of groupswho received physical practice Subsequently thestrategy acquired was arguably not appropriate for the

skill of the observers who were not able to plan andanticipate their movements to eyacute ciently perform thetracking task Kohl and Shea (1992) suggested ahierarchical control model to explain these reg ndingswhereby observational learning facilitates acquisition ata higher mental level but not at a muscle level Theselevels are somewhat similar to explicit and implicitmotor learning processes recently described by Gentile(1998) and also to Bernsteinrsquos (1996) higher (mental)level of action plans and lower level of sensory correc-tions (or synergies) These lower muscular levels arenecessary to ensure quality in the movement responsewhereas the higher level is required to shape plan andguide the movement

Verbal instruction and implicit and explicit learning

Gentile (1998) dithorn erentiated two learning processesthat are involved in the acquisition of motor skillsThese processes were termed èxplicitrsquo and implicitrsquo remacr ecting their dithorn erential accessibility to consciousawareness Although these terms have frequently beenused in the cognitive literature to explain both memoryand learning processes Gentile has provided a dereg -nition of these terms that is specireg c to motor learningExplicit processes are directed towards goal attainmentand developing the appropriate mapping between thelearner and the task Implicit processes concern theacquisition of the appropriate force-generation patternsthat enable eyacute cient movement production Pre-practiceinformation focuses the learner on being explicitlyaware of certain aspects of the task These aspects mayrelate to the movement response characteristics or tokey markers necessary for successful goal attainmentDuring the high jump the key marker is the end ofthe foot given that toe clearance is a requisite for asuccessful jump When performing a free-throw shotthe trajectory of the ball is the key marker and the pointabout which the performer should be most concerned(see Latash 1996)

Implicit and explicit processes are proposed to beinterdependent yet develop at dithorn erent rates Theexplicit process is expected to show a much faster rateof development than the implicit process Instructionsand demonstrations would be expected to inmacr uencethe development of the movementrsquos shape structure plusmnthe explicit process At this level the learner has anappreciation of the task requirements and a crudereference of the required behaviour Without experienceof the task however this explicit process may be insuyacute -cient to yield correct performance and learning Implicitappreciation of the underlying sensory experience ofthe movement will be needed Both processes need to betaken into account when assessing the benereg ts of pre-practice information An explicit representation of the


task possibly develops in parallel with practice experi-ence and does not need to precede correct production(cf Anderson 1992) Based on these proposals it islikely that tasks which place more emphasis on theperceptualplusmn motor demands and have complex responserequirements will benereg t less from explicit instructionabout the details of the movement at least early inacquisition

The results of several cognitive studies have alsoshown that an explicit representation or understandingof the task is not necessary for learning to take placeParticipants have been shown to be able to learn to con-trol complex interactive systems without an awarenessof the rules or regularities governing performance withlearning proceeding primarily in an implicit manner (egReber 1967 Berry and Broadbent 1984 Broadbentet al 1986) Similarly there is evidence that somemotor tasks may also be acquired implicitly Pew (1974)showed that individuals improved on a repeated seg-ment of a stimulus trace during a tracking task yet inpost-task testing showed little awareness that a segmentof the stimulus pattern had been repeated Lewicki et al(1992) have also shown that perceptual invariancescan be used to facilitate performance on simple key-response tasks (eg reaction times to repeated stimuli)without knowledge of these regularities Also in a com-puterized visualplusmn motor coordination task requiringjoystick responses to predictable and unpredictablestimuli Green and Flowers (1991) reported fewererrors for predictable stimuli in the absence of instruc-tion about this regularity When knowledge of stimuluscues was provided this explicit knowledge resulted inpoorer performance for participants who practised thetask under these conditions

The results of these studies show that where a strongperceptual component is present the detection ofinvariance or regularities in the stimulus array can bedetected even though the learner is not made aware ofthese regularities Indeed there is some suggestion plusmnespecially when the stimulusplusmn response regularities arenot provided on every trial (see Magill and Clark 1997)plusmn that explicitly directing individuals to search out thisinvariance might be detrimental for performance Theimportance of attention in mediating learning is criticalyet the complexity of attentional ethorn ects has to beappreciated Just because a variable or certain piece ofinformation might be related to task success explicitlydirecting participants to this information may hinderthe processing of additional important information orinterfere with the cognitive demands of the task

Not only have instructions been withheld during theacquisition of novel motor skills but implicit motorlearning has been examined by dividing attentionbetween the primary task and a cognitively demandingsecondary task This manipulation is believed to prevent

the acquisition of verbalizable rules and thus prevent anexplicit mode of learning Masters (1992) and Hardyet al (1996) compared the performance of an implicitlearning group a non-instructed control group andan explicit learning group in a golf-putting task Theexplicit learning group received specireg c instructions onhow to putt a golf ball These instructions were providedin between practice blocks and the participants wereencouraged to follow them as specireg cally as possibleThe implicit learning group showed improvement onthe task as a result of practice although they did notimprove at the same rate as the explicit learning andcontrol groups An interesting reg nding to emerge fromthese studies was that under conditions of stress onlythe implicit learning group continued to improve on thetask In fact the control participants and those wholearned under explicit learning conditions showeddecrements in performance

Masters (1992) compared the three groups afterpractice to assess the number of explicit rules acquiredabout golf putting He found that the control groupacquired more rules than the implicit learning groupThis led him to propose that the acquisition of explicitknowledge was the reason behind performance plateausor regressions under pressure More specireg callyMasters proposed that when performance is assessedunder pressure automatically produced actions orprocedures that are normally governed by lower-levelsystems get up-regulated to higher-order controlledsubsystems As such performance breaks down when`reinvested with explicit knowledgersquo (p 344)

More recently Maxwell et al (1999) found that thenumber of explicit rules acquired by the control groupin a similar golf-putting study correlated with scores ona reinvestment scale (Masters et al 1993) This scaleassessed a personrsquos predisposition to internalize actionsand focus on the mechanics of the movement The scaleproved to be a good predictor of performance such thatparticipants who scored highly performed poorly on agolf-putting task and were more likely to be rated as`chokersrsquo (ie to perform poorly under pressure) amonga group of university athletes

In a recent invited reaction to the implicit learningresearch of Masters Beek (2000) argued that implicitlearning should not be seen as an exceptional way oflearning motor skills (see also Seeger 1994) Rather forthe learning of perceptualplusmn motor skills explicit learningplusmn as dereg ned by the acquisition of verbalizable rulesand knowledge about the task plusmn was proposed to play aless signireg cant role than implicit learning As briemacr yremarked on earlier Bernstein (1996) also suggestedthat the motor system could be conceptualized by fourdithorn erent levels with various control responsibilities It isonly at the highest level of control (ie the action plans)that movement planning is critical and thus explicit


rules or knowledge are used Beek proposed thatsequencing skills involved in learning to dance or playmusic draw primarily on this high-level control whereasskills such as learning to ride a bike may be the functionof lower levels

Discovery learning

Attentional focus Although demonstrations are useful inlearning relatively simple motor tasks with high cogni-tive components little evidence shows that pre-practiceinformation facilitates the acquisition of more complexmovement tasks which place more demands on themotor components of the skill As with the study ofGreen and Flowers (1991) in the preceding sectionseveral investigators have shown that in the early stageof acquisition it may be best to leave participants todiscover for themselves how to attain the required goalsof the task Vereijken and Whiting (1989) and Whitinget al (1987) failed to observe benereg ts in performingski-like movements on a simulator after watching anexpert model Rather discovery learning participantsperformed as well as and on some variables better thana group who watched a skilled model (see also Andersonet al 1998) It was suggested that a demonstrationmight have caused the learners to be overly concernedwith the details of the movement and the achievementof movement form at the expense of goal attainmentSubsequently attention was divided between copyingthe movement and solving the motor task In the ski-simulator task discovery learners either had no or just afew preconceptions about the task their focus was onsolving the task and keeping the platform in motionAlthough dithorn erences in performance measures wereobserved across these learning groups Vereijken andWhiting (1989) pointed out that it is unclear whetherthe discovery learning groups approached the task insimilar ways to the instructed groups We have recentlyexamined this issue in investigations discussed in thereg nal sections of this paper (eg Hodges and Lee 1999)

Even though Bandura proposed that selectivelyattending to key features of a movement was criticalfor ethorn ective modelling in more complex movementresponse tasks directing attention to one of severalsources of information or components of the taskmight actually be detrimental to learning Den Brinkeret al (1986) provided learners with instructions andfeedback about one performance variable that related toperformance on the ski-simulator task that requiredfrequent macr uent and wide oscillating movementsImprovement on the attended variable was observed atthe expense of performance on other variables forexample improvement in frequency of the movementsat the expense of movement amplitude No discoverylearning control group was examined in this study

Subsequently van Emmerik et al (1989) conducteda follow-up experiment They found that discoverylearning was equally as ethorn ective as amplitude instruc-tion and even better than instruction and feedbackabout frequency and macr uency parameters Similar trade-othorn s in performance on one variable due to attention toother variables have been demonstrated by others (egSolley 1952 Maraj et al 1998)

In a more recent study rather than providing instruc-tion about a specireg c performance variable Wulf andWeigelt (1997) gave learners expert strategy instructionconcerning the appropriate time to exert force on theski-apparatus This variable was identireg ed in earlierstudies to be an important predictor of success (egVereijken et al 1992a) Therefore knowledge of whathad to be learned plusmn that is exploiting the reactive forcesof the system by varying phase lag plusmn informed as to thecontent for instructions without supposedly limitingattention to an isolated performance variable Despitethe relationship of this variable with success discoverylearners showed better acquisition and retention

Given the lack of practice for these learners beforeinstruction was received Wulf and Weigelt (1997)suggested that the expert strategy may have been in-appropriate until greater understanding and familiaritywith the task had been achieved However even after3 days of practice in a second experiment participantswere still unable to use this information ethorn ectively tofacilitate performance Rather almost all the partici-pants showed detriments when it was received asreduced movement frequency smaller amplitudemovements or increased jerkiness (ie more inhibitedmovements) The authors suggested that these detri-ments were due to a misdirection of attention promptedby the instructions This proposal has receivedconsiderable support (see below) These reg ndings alsoconcur with those of Masters (1992) who suggestedthat reinvesting well-practised actions with explicitknowledge changes the way a skill is performed

Wulf et al (1998 1999) have shown that instructionsthat direct attention to the limbs (eg the feet on a ski-apparatus the arms during a golf putt) rather thanthe external ethorn ects of the action (eg the wheels onthe ski-apparatus the golf club) are detrimental toacquisition The implication is that if movementdemonstrations or instructions directly or indirectlyencourage attention to the limbs learning is likely tosuthorn er Recently research has been conducted to try andprovide precise dereg nitions of external and internal focussuch that the theoretical nature of these ethorn ects can bedetermined as well as their practical implications Wulfet al (2000) found evidence to support the propositionthat attention should be directed to an ethorn ects-relevantcue plusmn that is the ball leaving rather than coming towardsthe racket in tennis Park et al (2000) found that a focus


on a far cue when learning to balance on a stabiliometerwas more benereg cial to retention performance than afocus on a near yet external cue No control groupswere examined in either of these studies however so itis diyacute cult to establish whether external-focus instruc-tions provide a benereg t beyond discovery learningconditions

The role of movement variability Similar to Wulf andWeigeltrsquos (1997) reg nding of more inhibited movementsas a result of attention-directing instructions Green andFlowers (1991) observed that participants who receivedinstruction throughout practice showed reduced joy-stick motion in all sessions compared with non-instructed participants Given that reduction in themotion of the joystick occurred early in the trialsequence when predictive cues could occur Greenand Flowers argued that a dithorn erent movement strategywas encouraged by the cognitive demands placed onindividuals in the instruction group These demandssuch as recall and implementation of the probabilityinstructions subsequently inhibited performanceTherefore one of the ethorn ects of movement-relatedinstructions or demonstrations may be a reduction inearly movement attempts at the desired action plusmn or atleast inhibited movements plusmn and a reduced explorationof the task environment Early in skill acquisition thesejerky-type movements are common observations inskills such as skiing gymnastics and dance Thinkingtoo much about aspects of the movement such as theappropriate time to bend the legs when performing aturn when skiing is likely to make performance morebroken and stuttered (see also Wulf and Weigelt 1997)A reduction in movement experience would be a par-ticular disadvantage if a novel movement was requiredthat was not already part of an existing movementrepertoire Without experience of the task dynamics andvaried stimulusplusmn response pairings the learner wouldexperience diyacute culties discovering the required co-ordination pattern

Vereijken (1991) argued that discovery learning couldwork positively compared with instructional learningby encouraging the discovery of optimal solutionsspecireg cally suited for the individual Allowing the indi-vidual to gain experience of other less than optimal solu-tions was proposed to enhance the learnerrsquos experienceof their environment and become aware of the con-straints acting upon them Instructions or performancemodels may serve to constrain this search for optimalsolutions that may be specireg cally suited to theindividual

For complex motor tasks it has been shown that earlyin learning performers freezersquo seemingly redundantdegrees of freedom in the motor apparatus to ease theprocess of acquisition For example in learning to per-

form on the ski-apparatus Vereijken et al (1992b)showed that the movements of the joints in the legs werehighly coupled in the early stages of acquisition Afterpractice the couplings of these joints especially those ofthe ankle decreased Trying to copy the movementsof an expert model may hinder learning because per-formers may not have suyacute cient control over theirdegrees of freedom to master the task in the same wayas an expert model Rather than solving the motorproblem by imposing their own constraints thengradually releasing them as control improves learnerswill try to copy the skilled model and might fail tobring together adequately the degrees of freedom in thesame way as the expert Paradoxically the instructionsor demonstrations would fail to constrain the motorsystem possibly by overwhelming the learner withinformation about the optimal control of many jointsyet constrain the search for solutions to the motorproblem that might be optimal for the learner at thatparticular time

Withholding movement instructions and demonstra-tions is only likely to be benereg cial if the learner tries toproduce new movements Langley and Simon (1981)argued that the learning mechanism must be able togenerate alternative behaviours if a teacher does notprovide them Trial-and-error learning can only bebenereg cial to learning if new behaviours are thenattempted This is where either intrinsic or extrinsicfeedback about the movement response plays a criticalrole in alerting the learner to incorrect performance andencouraging change on the next attempt Swinnen et al(1993) required participants to perform a three-partreversal movement with one arm and a simple macr exionmovement with the other yet start and end the move-ment of each arm simultaneously This task is diyacute cultto perform because of a preference to synchronize thelimbs Swinnen et al found that outcome informationwas equally as ethorn ective as more detailed knowledgeof performance specifying the exact displacements ofthe limbs Because both types of feedback alertedparticipants to a problem in coupling and thereforeencouraged the generation of new solutions the amountof detail or information supplied in the feedback wasunimportant

Recent learning experiments and pre-practice information

In several experiments we have examined the acquisi-tion of novel bi-manual coordination movements undervarious instructional manipulations As detailed earlierin bi-manual coordination two stable coordinationpatterns have been identireg ed plusmn in-phase and anti-phaseSome researchers (eg Swinnen et al 1991 Zanone andKelso 1992 1997 Lee et al 1995 Fontaine et al1997) have shown that trying to acquire a new relative


phase pattern such as 90deg relative phase is very dif-reg cult because of competition from these pre-existingbehaviours It is as though individuals get stuck in theearly stage of acquisition and need to engage in a pro-cess that allows them to break away from the attractionof competing behaviours Lee et al (1995) showed thatthe early stage of acquisition was characterized by initialstability in observed relative phase because early learn-ers showed a strong and stable bias towards the anti-phase and in-phase patterns After practice trials how-ever this stability was sacrireg ced and variabilityincreased due to within-trial switching between thesetwo stable patterns The authors argued that thisincrease in variability was indicative of the search todiscover the newly required movement pattern

Although learning novel bi-manual movements is stillrelatively simple compared with more real-world taskssuch as learning a backward somersault these bi-manual tasks require both spatial and temporal couplingand manifest themselves in a wide variety of leisurepursuits (eg catching juggling and playing a musicalinstrument) From a methodological standpointthey allow performance to be examined in a controlledlaboratory setting and permit quantireg cation of learningin comparison to existing behaviours Because mostindividuals cannot perform patterns other than in- andanti-phase pre-practice it is possible to examine the earlystage of skill acquisition where pre-practice informationin the form of instructions and movement demonstra-tions would be expected to have most inmacr uence

In one study (Hodges and Lee 1999) we manipu-lated instructions that informed individuals eithergenerally or specireg cally how to move their limbs to pro-duce the required 90deg pattern One group received astep-by-step specireg c guide as to the positioning of thehands every quarter of a cycle The general instruc-tion group received instructions specifying the startpositions of the hands (ie the spatial relationshipbetween the limbs) and the general rule that one handshould always lag the other by a quarter of a cycle Allparticipants received feedback at the end of each trial inthe form of a relative motion plot that depicted a circlepattern when the arms moved in the desired way In thisway it was possible to manipulate information abouthow to achieve the movement goal yet at the same timeproviding individuals with a task goal and associatedfeedback

No benereg ts for this type of instruction were foundRather both instruction groups showed detriments inacquisition retention or both The no-instructiongroup who received only feedback about circle pro-duction consistently showed the best performance Aspredicted this performance was associated with highwithin- and between-trial variability in relative phaseearly in practice enabling participants to break from

pre-existing stable movements Additionally a groupwho learned under secondary-task conditions failedto learn the new movement and showed low within-trial variability in relative phase because of consistentadherence to pre-existing movement patterns In reten-tion only the instructed participants were athorn ected bythe secondary-task conditions Arguably the cognitivedemands of the secondary task interfered with the pro-cesses required to perform the newly acquired coordin-ation pattern However that the instructed participantsgenerally showed more variability in performance ofthe newly practised pattern (and therefore the acquiredpattern was not as stable) could also account for theseregressions

In follow-up experiments (Hodges and Franks 20002001) visual demonstrations were also provided pre-practice and presented at both slow and regular speedsAgain this pre-practice information failed to facilitatelearning and in some instances showed negative ethorn ectson measures of performance However manipulation ofattentional focus instructions coupled with movementdemonstrations mediated the ethorn ects of pre-practiceinformation in a similar way to that discussed earlier byWulf and colleagues Directing participants to focuson the feedback plusmn the external ethorn ects of the action plusmnwhen watching the demonstration and performing themovement reduced the negative ethorn ects associated withinstructional provision (Hodges and Franks 2000)Repeated attention-directing instructions during acqui-sition were also useful later in no-feedback retentiontests Arguably these instructions promoted attentionto other sources of information during practice or atleast encouraged ethorn ortful practice once the generalform of the movement had been acquired

We also found that the type of on-line feedback wasparticularly important in mediating the learning ethorn ects(Hodges and Franks 2001) Simple information whichwas indirectly related to the relationship between thelimbs plusmn the circle pattern plusmn helped constrain the twolimbs to act as a coordinated unit Providing feedbackabout the position of both limbs even though moredetailed and arguably more informative was detri-mental to learning Similar results were reported bySwinnen et al (1993) It was our contention thatincreasing the compatibility between the demonstra-tions and the feedback would facilitate acquisition andthis limb feedback should have helped However thefeedback we provided was arguably too complex Videofeedback or perhaps a mirror would have providedlearners with more useful error information about thesimilarity of their movements to the pre-practice tem-plate We are currently exploring the benereg ts associatedwith coupling demonstrations and video feedbackand the implications of this method for enhancingerror-detection processes


Most recently we examined whether individual dif-ferences could play a role in mediating the ethorn ectsof instructions (Hodges and Franks in press) Anadditional purpose was to determine whether instruc-tions and demonstrations which showed how existingmovements could be adjusted to produce new move-ments would facilitate acquisition Swinnen (1996)has referred to this type of instruction as `transitionalinformationrsquo whereby learners are instructed as to whatto change to produce a new movement Although thisis a common method of instruction and there is goodreason to suppose that alerting participants to existingbiases would help them understand what was required(see Fitts and Posner 1967) it failed to facilitate per-formance This was the case irrespective of initial abilityAs in previous studies the non-instructed participantsgenerally outperformed the instructed groups Theinstruction groups showed decreased variability early inacquisition arguably owing to copying and avoidancebehaviour promoted by the instructions For exampleparticipants who were informed of a bias to the in-phasepattern and who were shown and instructed how toadjust this pattern to perform the required 90deg patternperformed worst overall On inspection of the relativephase data we discovered that participants in thisinstruction group showed a strong tendency to avoidin-phase movements Rather a bias to the anti-phasepattern was seen This failure to switch between thein- and anti-phase patterns early in acquisition wasdetrimental to the discovery of the newly requiredmovement

As a result of these studies we concluded that move-ment demonstrations and instructions relating to themovement of the limbs convey little useful informationin the early stage of acquisition if information aboutgoal attainment is available through feedback Instruc-tions were shown to hinder the break from pre-practicepatterns although there was some suggestion that theymay possibly help rereg ne the movement at a later stage ofpractice

Without examining all possible types of instructionit is impossible to make any reg rm conclusions abouttheir role in acquiring bi-manual skills or indeed skillacquisition in general Despite this limitation it is stillsurprising that instructions and demonstrations whichboth show and tell people how to perform a novelmovement and how to adjust existing movements fail toaid acquisition beyond that achieved from simple feed-back It is important to consider why this feedback wasso easy to use One reason was that it constrains the twolimbs to act as a single unit If we could identify instruc-tion that also acts to convey the general aspects of themovement then benereg ts to skill acquisition might beobserved When we have debriefed participants in thepast we have asked them to think about strategies they

used to help them perform They often report quiteabstract strategies such as `getting the feel for themovementrsquo or ìmagine being a circlersquo These insightswere similar to those reported by Franks and Stanley(1991) who found that participants learning to track astimulus waveform would hum or tap a rhythm to helpencode the visual pattern It is diyacute cult to imagine howthese could be used to help new learners Rather theseintuitions are probably only developed and understoodonce some expertise is achieved There is however aneed to examine more holistic methods of instructions(see also Swinnen 1996) that help the learner constrainthe degrees of freedom involved in the movement andfocus on more general aspects of the response We haveconsidered using the analogy of pulling on an elasticband to help participants incorporate the notion of aslight phase lag in the relation between the limbs but asyet these ideas await investigation

Summary

To summarize these research reg ndings and the empiricalstudies reviewed earlier we have constructed a macr owdiagram (see Fig 1) that illustrates how the processof skill acquisition is inmacr uenced by pre-practiceinformation

As depicted in the macr ow diagram the goals of the taskare a function of both the instructor and the learner Atdithorn erent stages of learning the performer will have adithorn erent understanding of these goals and thus theacquisition process will be athorn ected by the interaction ofthe two Earlier we discussed Fitts and Posnerrsquos (1967)ideas about transfer of habits and the importance ofexamining what skills the individual brings to the pro-cess Experience or lack of previous experience insimilar tasks will inmacr uence how the individual interpretseach set of circumstances and deals with the pre-practice information Masters et al (1993) showed howpersonality factors could impact on the skill acquisitionprocess and how pre-practice information should bedesigned to minimize the inmacr uence of undesirablethoughts or dispositions Performance goals that reduceattention to limb positioning and are more implicit maybe most suitable for individuals who tend to internalizeor reinvest actions

The goals specireg ed by the instructor will inmacr uencewhat behaviours are exhibited by the learner and thusinmacr uence the selection of an initial response A certainstrategy may be adopted as a result of this informationsuch as avoidance or adoption of a particular movementbehaviour Often demonstrations or limb-relatedinstructions will encourage the learner to seek outspecireg c feedback or error information that can becompared with pre-practice information or some sort of


Fig 1 Flow diagram to illustrate how the process of skill acquisition is inmacr uenced by various performance goals that serve toact upon the response both its selection and execution and associated processing of feedback The goals in turn are inmacr uencedby pre-existing movement skills and the results of practice attempts Postulated mechanisms underlying the ethorn ects ofvarious performance goals are suggested (dashed arrows) Both an explicit and implicit representation of the task is believed todevelop with practice with the implicit representation encompassing pre-existing movement habits and skills and the explicitrepresentation having a more direct inmacr uence on the choice and perception of the task goal KP = knowledge of performanceKR = knowledge of results

explicit transformation of this information Response-produced feedback may be augmented by the instructoror be a natural consequence of the movement (ieintrinsic) This feedback is critical for learning Withoutthe knowledge that an error has been made the per-former will not be motivated to change their responseon the next trial and thus improve performance If parti-cipants are provided with limb information pre-practiceit should not be too surprising that this will be theinformation they seek out during or after the movement

If the performance goals are multifaceted or theinstructor provides both outcome goals and movementgoals then some sort of goal confusion could result(Gentile 1972) This would inmacr uence how learnersapproach the task and more importantly where theydirect attention during execution Directing attention toone performance variable may be at the expense ofperformance on other variables or focusing on how themovement should be performed may athorn ect attainment(see van Emmerik et al 1989) As Swinnen (1996) haspointed out outcome variables such as knowledge ofresults serve a goal-directing role helping the learnerform a reference-of-correctness As illustrated in Fig 1and demonstrated in our studies this information willhave a strong inmacr uence on the performerrsquos subsequentperformance goals

With practice the skills and knowledge of the learnerwill change hence the proposed template of the move-ment that develops with practice and forms part of thelearnerrsquos existing skills will change The explicit com-ponent of this template might be some sort of trans-formation of the demonstrations instructions andaugmented feedback at least early in practice Thiscomponent would be particularly suited to tasks thatplace high demands on cognitive or memory processesThe implicit component might remacr ect more behaviouralchanges that result from `getting a feelrsquo for the move-ment (see Gentile 1998) Learning to monitor and useethorn ectively proprioceptive input would arguably be mostethorn ective for developing a more rereg ned implicit tem-plate which is likely to benereg t tasks that place moredemands on the motor aspects and subtleties of themovement These components of the template are simi-lar to the high and low levels of control discussed earlier(see Bernstein 1996) In Fig 1 we have representedthis template as a circle overlapping with the existingskills As such it will inmacr uence and be inmacr uenced bythe performance goals and the emerging response

One of the consequences of trying to implementinstructions or copy movement demonstrations isincreased consistency in the attempted movement in theearly stages of acquisition This was demonstrated by us


in bi-manual coordination studies as well by researchersstudying discovery learning (eg Vereijken 1991) andoptimal templates (eg Newell et al 1990) Thereforeone of the ways pre-practice information athorn ects theacquisition process is through the encouragement ornon-encouragement of variability in the attemptedmovement from trial to trial Unless the learner hasreason to change what they are doing new movementswill not be attempted This variability is therefore highlydependent on error information and also the pre-existing skills of the learner

In our reg rst study (Hodges and Lee 1999) detailedtask instructions athorn ected later performance under cog-nitively demanding conditions The suggestion is that itis best to decrease cognitive demands during responseexecution and that written instructions might inadvert-ently increase these demands Bandura (1986) hassuggested that children are particularly susceptible tothe detrimental ethorn ects of high cognitive demandsbrought about by task instructions If the attentiondemands are high participants could be distractedfrom processing feedback ethorn ectively and breaking fromstable undesirable behaviours Green and Flowers(1991) and Wulf and Weigelt (1997) also suggestedthat the ethorn ects of movement-related instructions couldbe related to increased cognitive demands wherebytoo much thought about `what to dorsquo leads tojerky inhibited movements In this way the increasedcognitive demands may exert their inmacr uence throughdecreased response variability Indeed the members ofthe secondary-task learning group in our study (Hodgesand Lee 1999) showed little within-trial performancevariability and failed to break away from pre-existingbehavioural biases even though they received feedbackat the end of each trial

Conclusions

We have shown that the ethorn ectiveness of demonstrationsand instructions is dependent on the existing skills ofthe learner and relatedly on the type of tasks or skillsthat are taught When the movement response is notan existing part of the learnerrsquos movement repertoireinstructions and movement demonstrations whichspecify the correct or optimal method for performing askill have little learning benereg t Although these con-clusions may at reg rst seem somewhat negative what isimportant is not that instructions and movementdemonstrations should not be used to enhance learningbut rather when they are used the instructor andlearner should be cautious and consider what informa-tion is conveyed This would require an appreciationof the many and sometimes conmacr icting goals a motorskill may have and how attentional processes might be

directed to certain cues as a result of these goalsAdditionally an appreciation of the learnerrsquos existingskills and potential limits in understanding the pre-practice information would also help in predicting whatbehaviours might be tried or indeed avoided as a resultof instruction

To highlight some implications of this work forlearners and teachers of motor skills generally andperhaps stimulate future research to verify these con-clusions outside traditional laboratory settings weprovide several recommendations for coaching practiceThese recommendations are presented and have beenbroadly categorized based on the specireg cation of thetask goal and feedback skill of the performer move-ment variability and focus of attention

Implications of this research for teachingmotor skills

Feedback and specireg cation of the task goal

middot Coaches should provide clear pre-practice informa-tion about the task goal plusmn whether this is a particularmovement pattern or a particular outcome plusmn and pro-vide feedback that is appropriate to the instructedgoal

middot Providing error information is vitally important earlyin skill acquisition It encourages athletes to trynew behaviours on their next attempt If no newbehaviours are attempted on the basis of intrinsicfeedback alone then some directed instruction maybe necessary

middot Feedback relating to the movement should be assimple as possible and convey important informationabout goal attainment This feedback should becompatible with the pre-practice information suchthat error information is easily attainable to deter-mine goal achievement and ethorn ective implementationof the pre-practice information

Skill

middot Although there are usually general features of amovement that dereg ne skilled performance optimalmovement templates do not always generalize wellacross individuals Individually based templatesmay be more appropriate in rereg ning and achievingconsistency in a skill

middot Demonstrations help to highlight the ordering of taskcomponents but not necessarily the teaching of themovement components If the necessary componentsto perform the action have already been acquiredthen demonstrations which facilitate the `stringing


togetherrsquo of these components are likely to be usefullearning aids

middot Physical limitations as a result of experience ordevelopment may limit the ability of the learnerto understand and use ethorn ectively pre-practiceinformation

middot Coaches should not assume experts and novicesare dealing with the same motor problem Demon-strations and instructions for experts are not neces-sarily the same as for novices Skill will athorn ect whatinformation an athlete is able to extract from ademonstration

Variability

middot Coaches should allow (and expect) error and vari-ability early on in learning to facilitate the discoveryof new movements

middot Demonstrations that are varied in content arerecommended to engage observers in the learningprocess and to inform them of possible solutions tothe motor problem

middot Early in acquisition experience of the task environ-ment should be encouraged Withholding detailedpre-practice information at this point may help toencourage this early exploration and hence vari-ability as long as suyacute cient error information isavailable

middot Instructions that build upon stable pre-existingmovement behaviours early in practice do notfacilitate the acquisition of skills that are not part ofthe movement repertoire This type of instructioncan promote avoidance behaviour and decrease earlyvariability necessary to break from undesirablemovements

Attentional focus

middot Individuals learn to respond ethorn ectively to certainaspects of the skill without being aware of it and with-out being instructed to learn Direct instruction orintervention is not always necessary to encouragelearning Additionally learners may be unable toverbalize the knowledge they have acquired despiteshowing improved performance on the task

middot Instructions and feedback direct attentional focus Ifperformance is to be assessed by more than one per-formance measure such as amplitude and frequencyor the task is so complex that directing attentionto one specireg c information source may be at thedetriment of other important factors then specireg cinstructions should be avoided

middot When instructions or demonstrations are provided

attention should be directed towards the externalethorn ects of the action rather than the limbs

middot If demonstrations are provided to convey the goalof the task this pre-practice information should beaccompanied by instructions that encourage atten-tion onto the feedback and the ethorn ects of the action onthe environment This additional instruction wouldalso promote more ethorn ortful practice and attention toinformation sources that may otherwise go unnoticedor be poorly processed

middot Some individuals may be predisposed to focusattention onto the mechanics of the movementthus hindering acquisition For these individuals theamount of explicit instruction pre-practice should bekept to a minimum

Acknowledgement

Funding for this paper was provided by a Social Science andHumanities Research Council of Canada grant awarded toIMF

References

Ackerman PL (1988) Determinants of individual dithorn er-ences during skill acquisition cognitive abilities andinformation processing Journal of Experimental Psychology

General 117 288plusmn 318Ackerman PL (1992) Predicting individual dithorn erences in

complex skill acquisition dynamics of ability determinantsJournal of Applied Psychology 77 598plusmn 614

Ackerman PL and Cianciolo AT (2000) Cognitiveperceptual-speed and psychomotor determinants ofindividual dithorn erences during skill acquisition Journal of

Experimental Psychology Applied 6 259plusmn 290Adams JA (1971) A closed-loop theory of motor learning

Journal of Motor Behavior 3 111plusmn 150Anderson DI Dialameh N Hilligan P Wong K and

Wong R (1998) Learning a slalom-ski-simulator taskwith a template of correct performance and concurrent orterminal videotape feedback Journal of Sport and Exercise

Psychology 20 S69Anderson JR (1982) Acquisition of cognitive skill

Psychological Review 89 369plusmn 406Arutyunyan GA Gurreg nkel VS and Mirskii ML

(1968) Investigation of aiming at a target Biophysics 13536plusmn 538

Bandura A (1969) Principles of Behavior Modireg cation NewYork Holt Rinehart amp Winston

Bandura A (1977) Social Learning Theory Englewood Clithorn sNJ Prentice-Hall

Bandura A (1986) Social Foundations of Thought and

Action A Social Cognitive Theory Englewood Clithorn s NJPrentice-Hall

Bandura A (1997) Self-Eyacute cacy The Exercise of Control NewYork Freeman


Beek PJ (2000) Toward a theory of implicit learning in theperceptual motor domain International Journal of Sport

Psychology 31 547plusmn 554Bernstein NA (1967) The Coordination and Regulation of

Movements London PergamonBernstein NA (1996) On dexterity and its development In

Dexterity and its Development (edited by ML Latash andMT Turvey) pp 9plusmn 235 Hillsdale NJ LawrenceErlbaum

Berry DC and Broadbent DE (1984) On the relationshipbetween task performance and associated verbalizableknowledge Quarterly Journal of Experimental Psychology 36209plusmn 231

Blake BG (1998) A guide to making practical demonstra-tions more ethorn ective in physical education lessons Bulletin of

Physical Education 34 59plusmn 64Brisson TA and Alain C (1996) Should common optimal

movement patterns be identireg ed as the criterion to beachieved Journal of Motor Behavior 28 211plusmn 223

Broadbent DE Fitzgerald P and Broadbent MH (1986)Implicit and explicit knowledge in the control of complexsystems British Journal of Psychology 77 33plusmn 50

Carroll WR and Bandura A (1982) The role of visualmonitoring in observational learning of action patternsmaking the unobservable observable Journal of Motor

Behavior 14 153plusmn 167Carroll WR and Bandura A (1985) Role of timing of

visual monitoring and motor rehearsal in observationallearning of action patterns Journal of Motor Behavior 17269plusmn 281

Carroll WR and Bandura A (1987) Translating cognitioninto action the role of visual guidance in observationallearning Journal of Motor Behavior 19 385plusmn 398

Davids K Handford C and Williams M (1994) Thenatural physical alternative to cognitive theories of motorbehaviour an invitation for interdisciplinary research insports science Journal of Sports Sciences 12 495plusmn 528

Deakin JM (1987) Cognitive components of skill inreg gure skating Dissertation Abstracts International 48 (5-B)1503

den Brinker BPLM Staumlbler JRLW Whiting HTA andvan Wierengen PCW (1986) The ethorn ect of manipulatingknowledge of results on the learning of slalom-type skimovements Ergonomics 29 31plusmn 40

Doody SG Bird AM and Ross D (1985) The ethorn ect ofauditory and visual models on acquisition of a timing skillHuman Movement Science 4 271plusmn 281

Feltz DL and Landers DM (1983) The ethorn ects of mentalpractice on motor skill learning and performance a meta-analysis Journal of Sport Psychology 5 25plusmn 57

Fitts PM and Posner MI (1967) Human PerformanceBelmont CA BrooksCole

Fleishman EA (1957) A comparative study of aptitude pat-terns in unskilled and skilled psychomotor performancesJournal of Applied Psychology 41 263plusmn 272

Fontaine RJ Lee TD and Swinnen SP (1997) Learning anew bimanual coordination pattern reciprocal inmacr uences ofintrinsic and to-be-learned patterns Canadian Journal of

Experimental Psychology 51 1plusmn 12Fowler C and Turvey MT (1978) Skill acquisition an

event approach with special reference to searching for theoptimum of a function of several variables In Information

Processing in Motor Control and Learning (edited byG Stelmach) pp 1plusmn 40 New York Academic Press

Franks IM (1997) The use of feedback by coaches andplayers In Science and Football III (edited by T ReillyJ Bangsbo and M Hughes) pp 267plusmn 278 London E amp FNSpon

Franks IM and Stanley ML (1991) Learning theinvariants of a perceptual motor skill Canadian Journal of

Psychology 45 303plusmn 320Franks IM and Wilberg RB (1982) The generation of

movement patterns during the acquisition of a pursuittracking task Human Movement Science 1 251plusmn 272

Gentile AM (1972) A working model of skill acquisition toteaching Quest 17 3plusmn 23

Gentile AM (1998) Implicit and explicit processes duringacquisition of functional skills Scandinavian Journal of

Occupational Therapy 5 7plusmn 16Gibson JJ (1979) The Ecological Approach to Visual

Perception Boston MA Houghton-Miuuml inGoulet C Bard C and Fleury M (1989) Expertise

dithorn erences in preparing to return a tennis serve a visualinformation processing approach Journal of Sport and

Exercise Psychology 11 382plusmn 398Green TD and Flowers JH (1991) Implicit versus explicit

learning processes in a probabilistic continuous reg ne-motorcatching task Journal of Motor Behavior 23 293plusmn 300

Haken H Kelso JAS and Bunz H (1985) A theoreticalmodel of phase transitions in human hand movementsBiological Cybernetics 51 347plusmn 356

Hardy L Mullen R and Jones G (1996) Knowledge andconscious control of motor actions under stress British

Journal of Psychology 87 621plusmn 636Higgins S (1991) Motor skill acquisition Physical Therapy

71 123plusmn 139Hodges NJ and Franks IM (2000) Focus of attention

and coordination bias implications for learning a novelbimanual task Human Movement Science 19 843plusmn 867

Hodges NJ and Franks IM (2001) Learning a co-ordination skill interactive ethorn ects of instruction and feed-back Research Quarterly for Exercise and Sport 72 132plusmn 142

Hodges NJ and Franks IM (in press) Learning as a func-tion of coordination bias building upon pre-practicebehaviours Human Movement Science

Hodges NJ and Lee TD (1999) The role of augmentedinformation prior to learning a bimanual visualplusmn motorcoordination task do instructions of the movement patternfacilitate learning relative to discovery learning British

Journal of Psychology 90 389plusmn 403Howard JH Jr Mutter SA and Howard DV (1992)

Serial pattern learning by event observation Journal of

Experimental Psychology Learning Memory and Cognition18 1029plusmn 1039

Hughes M and Franks IM (1997) Notational Analysis in

Sport London E amp FN SponKelso JAS (1984) Phase transitions and critical behavior

in human bimanual coordination American Journal of

Physiology Regulatory Integrative and ComparativePhysiology 15 R1000plusmn R1004


Kelso JAS (1995) Dynamic Patterns The Self-organization of

Brain and Behaviour Cambridge MA MIT PressKelso JAS Scholz JP and Schrdquoner GS (1986) Non-

equilibrium phase transitions in coordinated biologicalmotion critical uctuations Physics Letters A118 279plusmn 284

Kohl RM and Shea CH (1992) Pew (1966) revisitedacquisition of hierarchical control as a function ofobservational practice Journal of Motor Behavior 24247plusmn 260

Langley P and Simon HA (1981) The central role of learn-ing in cognition In Cognitive Skills and Their Acquisition

(edited by JR Anderson) pp 361plusmn 380 Hillsdale NJLawrence Erlbaum

Latash ML (1996) The Bernstein problem how does thecentral nervous system make its choices In Dexterity and its

Development (edited by ML Latash and MT Turvey)pp 277plusmn 303 Hillsdale NJ Lawrence Erlbaum

Lee TD and White MA (1990) Inmacr uence of an unskilledmodelrsquos practice schedule on observational motor learningHuman Movement Science 9 349plusmn 367

Lee TD Swanson LR and Hall AL (1991) What isrepeated in a repetition Ethorn ects of practice conditions onmotor skill acquisition Physical Therapy 71 150plusmn 156

Lee TD Swinnen SP and Verschueren S (1995) Relativephase alterations during bimanual skill acquisition Journal

of Motor Behavior 27 263plusmn 274Lee TD Blandin Y and Proteau L (1996) Ethorn ects of

task instructions and oscillation frequency on bimanualcoordination Psychological Research 59 100plusmn 106

Lewicki P Hill T and Czyzeska M (1992) Nonconsciousacquisition of information American Psychologist 47796plusmn 801

Magill RA and Clark R (1997) Implicit versus explicitlearning of pursuit-tracking patterns Journal of Sport and

Exercise Psychology 19 S85Maraj BKV Allard F and Elliott D (1998) The ethorn ect of

nonregulatory stimuli on the triple jump approach runResearch Quarterly for Exercise and Sport 69 129plusmn 135

Martens R Burwitz L and Zuckerman J (1976) Modelingethorn ects on motor performance Research Quarterly 47277plusmn 291

Masters RSW (1992) Knowledge knerves and know-howBritish Journal of Psychology 83 343plusmn 358

Masters RSW Polman RCJ and Hammond NV (1993)`Reinvestmentrsquo a dimension of personality implicated inskill breakdown under pressure Personality and Individual

Dithorn erences 14 655plusmn 666Maxwell J Masters R Eves F and MacMahon K (1999)

From novice to no know-how a longitudinal study ofimplicit motor learning Journal of Sports Sciences 17 608

McCullagh P and Caird JK (1990) Correct and learningmodels and the use of model knowledge of results in theacquisition and retention of a motor skill Journal of HumanMovement Studies 18 107plusmn 116

McCullagh P and Weiss MR (2000) Modeling con-siderations for motor skill performance and psychologicalresponses In Handbook of Sport Psychology 2nd edn (editedby R Singer) pp 205plusmn 238 New York Wiley

More KG and Franks IM (1996) Analysis and modireg ca-tion of verbal coaching behaviour the usefulness of a

data-driven intervention strategy Journal of Sports Sciences14 523plusmn 543

Newell KM (1981) Skill learning In Human Skills (editedby DH Holding) pp 203plusmn 225 New York Wiley

Newell KM (1986) Constraints on the development ofcoordination In Motor Development in Children Aspects of

Coordination and Control (edited by M Wade and HTAWhiting) pp 341plusmn 360 Dordrecht Martinus Nijhothorn

Newell KM Morris LR and Scully DM (1985)Augmented information and the acquisition of skill inphysical activity In Exercise and Sport Sciences ReviewsVol 13 (edited by RL Terjung) pp 235plusmn 261 LexingtonMA Collamore Press

Newell KM Carlton MJ and Antoniou A (1990) Theinteraction of criterion and feedback information inlearning a drawing task Journal of Motor Behavior 22 8plusmn 20

Park J-H Shea C McNevin N and Wulf G (2000)Attentional focus and the control of dynamic balanceJournal of Sport and Exercise Psychology 22 S85

Parker JF and Fleishman EA (1959) Prediction ofadvanced levels of proreg ciency in a complex tracking taskUSAF Wright Air Development Center Technical Report

59plusmn 255 vi 57Parker JF and Fleishman EA (1961) Use of analytical

information concerning task requirements to increase theethorn ectiveness of skill training Journal of Applied Psychology45 295plusmn 302

Pew RW (1974) Levels of analysis in motor control Brain

Research 71 393plusmn 400Reber AS (1967) Implicit learning of artireg cial grammars

Journal of Verbal Learning and Verbal Behavior 6 855plusmn 863Ross D Bird AM Doody SG and Zoeller M (1985)

Ethorn ect of modeling and video-tape feedback with knowledgeof results on motor performance Human Movement Science4 149plusmn 157

Schmidt RA (1975) A schema theory of discrete motor skilllearning Psychological Review 82 225plusmn 260

Schmidt RA (1976) Control processes in motor skillsExercise and Sport Sciences Reviews 4 229plusmn 261

Schmidt RA (1988) Motor Control and Learning A

Behavioral Emphasis 2nd edn Champaign IL HumanKinetics

Schmidt RA (2001) Motor schema theory after 26 yearssome remacr ections and future directions Journal of Sport and

Exercise Psychology 23 S4Schmidt RA and Lee TD (1998) Motor Control and

Learning A Behavioral Emphasis 3rd edn Champaign ILHuman Kinetics

Schmidt RC Carello C and Turvey MT (1990)Phase transitions and critical macr uctuations in the visual co-ordination of rhythmic movements between peopleJournal of Experimental Psychology Human Perception and

Performance 16 227plusmn 247Schrdquoner G Zanone PG and Kelso JAS (1992) Learning

as a change of coordination dynamics theory and experi-ment Journal of Motor Behavior 24 29plusmn 48

Scully DM and Newell KM (1985) Observationallearning and the acquisition of motor skills toward a visualperception perspective Journal of Human Movement Studies11 169plusmn 186


Seeger CA (1994) Implicit learning Psychological Bulletin115 163plusmn 196

Shithorn rin RM and Schneider W (1977) Controlled andautomatic human information processing II Perceptuallearning automatic attending and a general theory Psycho-logical Review 84 127plusmn 190

Skinner BF (1953) Science and Human Behavior New YorkMacmillan

Solley WH (1952) The ethorn ects of verbal instruction of speedand accuracy upon the learning of a motor skill Research

Quarterly 23 231plusmn 240Starkes JL Deakin JL Lindley S and Crisp F (1987)

Motor versus verbal recall of ballet sequences by youngexpert dancers Journal of Sport Psychology 9 222plusmn 230

Swinnen SP (1996) Information feedback for motor skilllearning a review In Advances in Motor Learning and Control

(edited by HN Zelaznik) pp 37plusmn 66 Champaign ILHuman Kinetics

Swinnen SP De Pooter A and Delrue S (1991)Moving away from the in-phase attractor during bimanualoscillations In Studies in Perception and Action (edited byPJ Beek RJ Bootsma and PCW van Wieringen)pp 315plusmn 319 Amsterdam Rodopi

Swinnen SP Walter CB Lee TD and Serrien D(1993) Acquiring bimanual skills contrasting forms ofinformation feedback for interlimb decoupling Journal of


Thorndike EL (1927) The law of ethorn ect American Journalof Psychology 39 212plusmn 222

Van Emmerik REA den Brinker BPLM Vereijken Band Whiting HTA (1989) Preferred tempo in thelearning of a gross cyclical action Quarterly Journal of

Experimental Psychology 41A 251plusmn 262Vereijken B (1991) The Dynamics of Skill Acquisition Meppel

Kripps ReproVereijken B and Whiting HTA (1989) In defense of dis-

covery learning In Catching Up Selected Essays of HTA

Whiting (edited by PCW van Wieringen and RJBootsma) pp 155plusmn 169 Amsterdam Free University Press

Vereijken B Whiting HTA and Beek WJ (1992a) Adynamical systems approach to skill acquisition Quarterly

Journal of Experimental Psychology 45A 323plusmn 344Vereijken B Whiting HTA Newell KM and van

Emmerik REA (1992b) Free(z)ing degrees offreedom in skill acquisition Journal of Motor Behavior 24133plusmn 142

Vogt S (1996) The concept of event generation in movementimitation plusmn neural and behavioral aspects Corpus Psyche et

Societas 3 119plusmn 132Whiting HTA and den Brinker BPLM (1982) Image of

the act In Theory and Research in Learning Disabilities

(edited by JP Das RF Mulcahy and AE Wall)pp 217plusmn 235 New York Plenum Press

Whiting HTA Bijard MJ and den Brinker BPLM(1987) The ethorn ect of the availability of a dynamic modelon the acquisition of a complex cyclical action QuarterlyJournal of Experimental Psychology 39A 43plusmn 59

Wulf G and Weigelt C (1997) Instructions about physicalprinciples in learning a complex motor skill to tell ornot to tell Research Quarterly for Exercise and Sport 68362plusmn 367

Wulf G Hrdquoss M and Prinz W (1998) Instructions formotor learning dithorn erential ethorn ects of internal versusexternal focus of attention Journal of Motor Behavior 30169plusmn 179

Wulf G Lauterbach B and Toole T (1999) Learningadvantages of an external focus of attention in golf Research

Quarterly for Exercise and Sport 70 120plusmn 126Wulf G McNevin NH Fuchs T Ritter F and Toole T

(2000) Attentional focus in complex skill learning Research

Quarterly for Exercise and Sport 71 229plusmn 239Zanone P-G and Kelso JAS (1992) Evolution of

behavioral attractors with learning nonequilibriumphase transitions Journal of Experimental Psychology HumanPerception and Performance 18 403plusmn 421

Zanone P-G and Kelso JAS (1997) Coordinationdynamics of learning and transfer collective and com-ponent levels Journal of Experimental Psychology Human

Perception and Performance 23 1454plusmn 1480


instruction and demonstration Despite the shortage ofunequivocal empirical evidence to support the provisionof pre-practice information most coaching educationprogrammes sanction its use and consider it to be a keyfactor in teaching motor skills A physical education(PE) researcher specialist commented Ìt would be astrange PE lesson indeed it would be a rare PE lessonif some form of physical demonstration was not shownto some or all of the pupilsrsquo (Blake 1998 59)

In fact little scientireg c consensus exists about the rolepre-practice information plays in the learning processbeyond the sometimes necessary role of relaying to thelearner the goals of the task This gap in knowledge isparticularly interesting given that augmented feedbackduring and after skill execution has received markedattention in the motor learning literature In their mostrecent motor control and learning textbook Schmidtand Lee (1998) echo these sentiments by concludingtheir section on pre-practice conditions with thefollowing statement few investigations have beencarried out on the nature of instructions and muchwork remains to be done in assessing their value inlearningrsquo (p 289)

In this paper we review research that has focusedon the ethorn ectiveness of pre-practice information forlearning motor skills and suggest some guidelines for`best practicersquo coaching behaviour that is based on thisempirical evidence First we discuss the theory under-lying these research reg ndings and the implications ofthese theories for providing pre-practice informationWe include a discussion of traditional motor learningtheories (eg Adams 1971 Schmidt 1976) and otherapproaches that posit a primary role for movementrepresentations (eg Fitts and Posner 1967 Bandura1986) Additionally we discuss less prescriptiveapproaches with an emphasis on the constraints in themotor system that are responsible for the emergingmovement

We then review empirical reg ndings and contrastexplicit learning strategies with more implicit discoverylearning environments Dithorn erent performance goals areexamined such as criterion templates demonstrationsand verbal instructions and we try to determine themechanisms underlying these various instructionalmethods including focus of attention and movementvariability These various performance goals how theyinmacr uence the emerging movement and in turn areinmacr uenced by pre-existing skills and the results of themovement are subsequently illustrated in a macr ow dia-gram (see Fig 1) In the reg nal sections we present aseries of experiments from our laboratory in which weexamined the ethorn ects of dithorn erent instructional variableson the learning of novel bi-manual coordinationmovements We discuss how various performancegoals athorn ect the selection of a movement response and

the subsequent processing of feedback We end with atable that outlines some of the practical implications ofthe research reviewed for the teaching of motor skills

Motor learning theory and implications forinstruction and demonstration

Classic theories of skill acquisition

The question of how people acquire novel motor skillshas long been a topic of interest This dates back to theearly work of the behaviourists (eg Thorndike 1927Skinner 1953) when the outcome of the movementand the determination of environmental contingencieswere of primary signireg cance With the cognitiverevolution came an associated interest in the processesresponsible for learning new behaviours and conceptsrelated to information processing In motor behaviourtwo theories have received most attention These areAdamsrsquo (1971) closed-loop theory of motor controlhighly inmacr uenced by the reg eld of cybernetics andSchmidtrsquos (1975 1976 1988) schema theory wherebymotor programs are proposed to be responsible formovement execution much like software programs incomputers

Schmidt and Adams proposed that learning results inthe development of two representations of the move-ment in memory either actual memory traces (Adams1971) or abstractions called schemasrsquo (Schmidt 1975)These representations are responsible for initiating themovement and evaluating its ethorn ectiveness both duringand after execution The perceptual trace or recognitionschema is responsible for evaluating the ethorn ectivenessof the response and therefore acts as a referencemechanism to compare the actual sensory con-sequences with those desired After practice the per-ceptual trace represents those sensory or feedbackcharacteristics of the correct response whereas therecognition schema allows the learner to estimate theexpected sensory consequences of an action becausegeneral (invariant) features of the movement are storedGiven the importance of the detection and recognitionof errors in both Adamsrsquo and Schmidtrsquos theoriespre-practice information which is designed to rereg nethe movement template (trace or schema) should beimportant in aiding this process Such information thatallows the learner to extract easily and eyacute ciently theexpected response either in verbal or visual form wouldarguably aid the development of a reference with whichfeedback about the movement can be compared (seealso Swinnen 1996) This assumption only holdshowever if a movement template is there to begin

Pre-practice information might easily inmacr uence theresponse that is evoked or selected in the early stages of

































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References

























































































































































































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References
















































































































































































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References







































































































































































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References
































































































































































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References























































































































































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References













































































































































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References




































































































































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References



























































































































































Discovery learning





























Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References















































































































































































Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References






































































































































































Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References






























































































































































Summary













Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References


































































































































































Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References




























































































































































Conclusions









Skill







Variability





Attentional focus







Acknowledgement


References





























































































































































Variability





Attentional focus







Acknowledgement


References




























































































































































































































































































































































































































Documents

Hodges and Franks