Technical Report 431 (.,2

EFFECTS OF PRESENTATION- AND TEST-TRIAL TRAINING

ON MOTOR ACQUISITION AND RETENTION

Joseph D. Hagman

00O TRAINING TECHNICAL AREA

f U. S. Army<:1l Research Institute for the Behavioral and Social Sciences

I~J

January 1980

Approved for public release, distribution unlimited.

80 7 1-045 0

U. S. ARMY RESEARCH INSTITUTEFOR THE BEHAVIORAL AND SOCIAL SCIENCES

A Field Operating Agency under the Jurisdiction of theDeputy Chief of Staff for Personnel

FRANKLIN A. HARTJOSEPH ZEIDNER Colonel, US ArmyTechnical Director Commander

NOTICES

DISTRIBUTION Primarv distribution of this report hal been mde by ARI. Plean address correspondenceconcerning distribution of reports to. U. S. Army Rieerch Institute for the Behavioral and Social Sciences,ATTN: PERI.TP, 5001 Eisnhowar Avenue, Alexandria. Virginia 22333.

FiNAL ISPOIITION This report may be destroyed when it is no longer needed. Pleea do not return it tothe U. S. Army Research Institute for the Behavioral and Social Sciences.

P9.I The findings in this report are not to be construed as en official Depoartlmlent of the Army position.unless so designated by other authorized documents.

UnclassifiedSECURITY CLASSIFICATION OF THIS PAGE (W14hen Dole Entered)

REPOT DCUMNTATON AGEREAD INSTRUCTIONSREPOR DOCUENTATON PAE / 'AEFORE COMPLETING FORM6J 2. GOVT ACCESSION NO:

j Technical~g aert 43i n" 16- 1v_

FFECTS OF I RESENTATION- AND TEST-TRIAL TRAININGPN OTOR ACQUISITION AND RETENTION*

6. PERFORMING ORG. REPORT NUMBER

7. AUTHOR(&) 8. CONTRACT OR GRANT NUMBER(e)

Josep D./Hagman

9. PERFORMING ORGANIZATION NAME AND ADDRESS to 8. PIROGRAM ELEMENT PROJECT, TASKUAryResearch Institute for the Behavirl&*~ OKUI UBR

Social Sciences, PERI-IIl1- 2Tl6l10(lA9lB5001 Eisenhower Avenue, Alexandria, VA 22333

It. CONTROLLING OFFICE NAME AND ADDRESS Ri EPO.RLGAZS..Deputy Chief of Staff for Personnel _, raw W 80Washington, DC 20310 "T. 3, WJfUWWROF -

2214. MONITORING AGENCY NAME & ADDRESSI different from Controling Off ice) 15, SECURITY SLAII47fof .rt r

UnclassifiedIS15. DECL ASSI FIC ATION/0DOWNGRADING

'I SCHEDULE

16. DISTRIBUTION STATEMENT (of this Report)

Approved for public release; distribution unlimited.

17. DISTRIBUTION STATEMENT (of the abstract entered in Block 20, If different from Report,,

IS. SUPPLEMENTARY NOTES

19. KEY WORDS (Continue on fev'ers* side if necessary and identify by block numiber)

Training RepetitionMotor Skills Presentation TrialsLearning Test TrialsRetention Linear Positioning

20. ABSTRACT (Contnde on reverse side Of necessary and identify by block numnber)Thi o~ari.-t1examined the relative effects of three different training methods

on the acquisition and retention of a positioning motor task. Three independentgroups of subjects (N = 15 per group) performed three training trial cycles con-sisting of six trials each. Training methods differed in their emphasis onpresentation (p) and test (t) trials during each cycle. For one group, a cycleconsisted of three p- and t-trials administered in alternation. For anothergroup, the first five trials of each cycle were p-trials and the sixth was a -

D 1AN73 1473 EDITION OF I NOV 65IS5OBSOLETE Unclassified

SECURITY CLASSIFICATION OF THIS PAGE (When DUU.d)

Unclassified

SECURITY CLASSIFICATION OF THIS PAGO(enWJ Da4ta Etemd

t-trial. For the last group, the first trial was a p-trial and the next fivewere t-trials. Group acquisition performance was compared at the last trialof each cycle while retention was compared 3 mn. and 24 hr. after acquisition.

Absolute error scores indicated that acquisition and short-term retention werebest when training emphasized p- and t-trial alternation and p-trial repetitionwithin cycles, whereas long-term retention was best when training stressedt-trial repetition. Results suggest that testing is an effective way to enhancelong-term retention of motor skill. This enhancement could be realized bychanging the emphasis of training from presentation to testing without addedexpenditures in training time, money and personnel. If, instead of long-termretention, the goal is rapid acquisition and short-term retention, trainingmethods which emphasize either alternation of presentation and testing orrepeated presentation would be most effective.

j.For

/TI

~!

Un class if~ *i tCUfITy CLASSIFICATION OF THIS PAGEtIRU DUE. NLIIem

Technical Report 431

EFFECTS OF PRESENTION- AND TEST-TRIAL TRAINING

ON MOTOR ACQUISITION AND RETENTION

Joseph D. Hagman

Submitted by.Milton S. Katz, Chief

TRAINING TECHNICAL AREA

Approved By:

E. Ralph DusekPERSONNEL AND TRAININGRESEARCH LABORATORY

U.S. ARMY RESEARCH INSTITUTE FOR THE BEHAVIORAL AND SOCIAL SCIENCES5001 Eisenhower Avenue, Alexandria, Virginia 22333

Office, Deputy Chief of Staff for PersonnelDepartment of the Army

January 1980

Army Project Number Motor Skills Training2T1611O1A91B

04red for oublic n1gm: distribution~ unlimited.

ARI Research Reports and Technical Reports are intended for sponsors ofR&D tasks and for other research and military agencies. Any findings readyfor implementation at the time of publication are presented in the last partof the Brief. Upon completion of a major phase of the task, formal recom-mendations for official action normally are conveyed to appropriate militaryagencies by briefing or Disposition Form.

iv

. .".. ...

FOREWORD

The Training Technical Area of the Army Research Institute for theBehavioral and Social Sciences (ARI) conducts research in support of thesystems engineering concept of training. A major objective of thisresearch is to develop the fundamental data and technology necessary tofield integrated systems for improving individual job performance. Suchsystems include Skill Qualification Testing (SQT), job performance aids,and training courses both in schools and in the field.

This report is one of a series on specific topics in the area ofskill acquisition and retention. In response to requirements by theDeputy Chief of Staff for Training of the Army Training and DoctrineCommand (TRADOC), the long-term research goal is to develop methods forpredicting proficiency loss for all types of skills and for determiningeffective training procedures for reducing this loss. The present workrepresents a basic research effort completed by ARI personnel under ArmyProject 2TI611OIAl9B.

(JO EPH ZE ER'Shnical Director

v

A V

EFFECTS OF PRESENTATION- AND TEST-TRIAL TRAINING ON MOTOR ACQUISITIONAND RETENTION

BRIEF

Requirement:

To evaluate the relative effectiveness of three motor task trainingmethods which differ in their emphasis on presentation and test trials.

Procedure:

Tnree groups of 15 participants received 18 training trials on asimple motor task. The 18 training trials were divided into threecycles of six trials, containing both presentation and test trials.During presentation trials, participants studied the criterion movementto be learned by moving a sliding mechanism along a linear track for adistance of 250 mm before contacting a mechanical stop. During testtrials, they tried to recall the criterion movement by moving the slidefor the same distance without the aid of the mechanical stop.

In the acquisition part of the experiment, the sequence of presentationand test trials performed within cycles differed for each traininggroup. For the STANDARD group, a cycle consisted of three presentationand three test trials administered in alternation. For the PRESENTATIONgroup, the first five trials of each cycle were presentation trials andthe sixth was a test trial. For the TEST group, the first trial was apresentation trial and the next five were test trials. In the retentionpart of the experiment, all participants performed a single test trialat both 3 minutes and 24 hours after the last training trial.

Findings:

The three training methods had different effects on acquisition andretention. Absolute (unsigned) error scores indicated that acquisitionand short-term (3 minute) retention were best for the STANDARD andPRESENTATION groups, whole long-term (24 hour) retention was best forthe TEST group.

Utilization of Findings:

Testing is an effective way to enhance long-term-retention of motorskills. This enhancement can be achieved by changing the emphasis intraining from presentation to testing, without the need for additionaltraining time, money or personnel. If instead of long-term retention,the goal is rapid acquisition and short-term retention, training which

vii I -

p" NA NOT1UL

emphasizes either alternation of presentation and testing or repeatedpresentation would be most effective. Additional research is needed todetermine how well these laboratory results will generalize to actualmilitary motor skills.

viii

EFFECTS OF PRESENTATION- AND TEST-TRIAL TRAINING ON MOTOR ACQUISITIONAND RETENTION

CONTENTS

Page

INTRODUCTION...........................

METHOD............................3

Subjects..........................3Apparatus. .......................... .. .. ...... 4Design...........................4Procedure.........................5

RESULTS...........................5

Acquisition........................7Retention........................10

DISCUSSION...........................11

CONCLUSIONS.........................14

REFERENCES...........................17

DISTRIBUTION.........................21

ix

LIST OF FIGURES

Page

Figure 1. Trial sequence for each training group(P = Presentation; T = Test) .. ............... 6

2. Mean algebraic error resulting from t-trialexecution by the STANDARD, PRESENTATION andTEST training groups at acquisition andretention ........ ........................ 8

3. Mean absolute error resulting from t-trialexecution by the STANDARD, PRESENTATION andTEST training groups at acquisition and

retention ........ ....................... 9

x

EFFECTS OF PRESENTATION- AND TEST-TRIAL TRAINING ON MOTOR ACQUISITION

AND RETENTION

INTRODUCTION

A long-term goal of the Army is the development of effective methodsfor training all types of Army-related skills. Of constant interest isthe question of which specific training methods promote the highestlevels of skill acquisition and retention. Much of the theoretical andempirical information relating to this question has been derived fromthe conduct of basic research experiments in the areas of verbal andmotor learning. In these experiments, training has involved the execu-tion of both presentation (p) and test (t) trials. During p-trials,subjects study information to be learned and during t-trials they attemptto recall it from memory. The number and sequential arrangement of p-and t-trials performed during training has depended on the particularmethod adopted. The standard training has involved the alternation ofp- and t-trials (e.g., Tulving, 1967; Wrisberg & Schmidt, 1975) butother methods which emphasize either p- or t-trial repetition also havebeen used (Adams & Dijkstra, 1966; Bilodeau & Bilodeau, 1958; Hogan &Kintsch, 1971).

The Army's question of which training method most effectively promotesacquisition and retention is difficult to answer because relevant theoriesmake conflicting predictions regarding the role of p- and t-trials. Forexample, from a traditional learning theory viewpoint, where p-trialsare seen as having an effect similar to reinforcement, training methodswhich repeat p-trials should be more effective than those which repeatt-trials. Repetition of t-trials reduces the number of reinforcementopportunities, and therefore, should retard both acquisition and retention.From a contemporary cognitive viewpoint, however, information processing

activities such as recall and internal generation of to-be-learned itemsare considered important aspects of acquisition and retention (Bjork,1975; Dosher & Russo, 1976). Because t-trials provide an opportunity toperform these activities on the information studied during p-trials,training methods which repeat t-trials should be more effective thanthose which repeat p-trials.

A look at empirical evidence related to the issue of which trainingmethods are most effective also reveals inconsistencies. Most of thisevidence has come from research in verbal task learning. Here, therelative effects of p- and t-trial repetition during training have beenof interest for a long time (Gates, 1917; Hellyer, 1962; Raffel, 1934)but have only recently received systematic investigation. In general,investigators have found that training methods which emphasize p-trialrepetition produce superior acquisition (Hogan & Kintsch, 1971; Thompson,Wenger & Bartling, 1978, Exp III) whereas those which emphasize t-trialrepetition produce superior retention (Allen, Mahler & Estes, 1969;Hogan & Kintsch, 1971; Rosner, 1970; Thompson, et al., 1978, Exp II;)

1]

In contrast, the pattern of results for motor task learning has beensomewhat different. Generally, emphasis on p-trial repetition duringtraining has enhanced both acquisition (Holding & Macrae, 1964) andretention (Adams & Dijkstra, 1966), but emphasis on t-trial repetitionhas not. Although both subjective recall consistency and error detectionability have developed as a function of repeated t-trials (Newell, 1974;Seashore & Bevelas, 1941) movement accuracy has not been found to improve(Holding & Macrae, 1964; Newell, 1976, Exp I; Thorndike, 1927) exceptafter considerable prior t-tr il repetition (Newell, 1976, Exp I andIII). In fact, accuracy typically has decreased during both acquisitionand retention when t-trials have been repeated during training (Bilodeau& Bilodeau, 1958; Duffy, Montague & Laabs, 1975).

Thus, training methods stressing p-trial repetition have had relativelyconsistent beneficial effects on both motor and verbal task performance,especially at acquisition. In contrast, methods stressing t-trials havepositively affected verbal task retention but have negatively affectedboth motor task retention and acquisition. Specific reasons for thisdifferential effect of t-trial repetition on verbal and motor taskperformance are difficult to pinpoint because of the many differencesthat exist between the two areas of research. One suggested reason,however, centers around the difference in information processing activityrequired of subjects at p- and t-trials during motor and verbal tasktraining. In verbal task training, p- and t-trials are procedurallydistinct and require dissimilar information processing activities.During p-trials, items to be learned are shown to subjects for study.During t-trials, these items are removed and subjects are required torecall them from memory. In motor task training, however, p- and t-trials are procedurally similar and require very similar informationprocessing activities. At p-trials, subjects attempt to recall acriterion movement from memory. This recall attempt is followed byknowledge of results regarding recall accuracy, typically in visual orverbal form. At t-trials, subjects are also required to recall thecriterion movement but knowledge of results is not provided. Thus,motor task training requires recall at the execution of both p- and t-trials whereas verbal task training requires study at p-trials andrecall at t-trials. To the extent that study and recall processes havebeen found to differentially affect verbal acquisition and retention(e.g., Hogan & Kintsch, 1971), it is necessary to create a commonprocedural environment in motor task training such that process effectson motor acquisition and retention can also be examined.

Another suggested reason for the different effects of t-trialrepetition on verbal and motor task performance stems from the differencein retention interval lengths used to investigate the retention of eachtype of task. For example, the effects of repeated t-trial training onverbal task retention have been examined primarily using long-termretention intervals (e.g., Allen, et al., 1969) whereas short-term

2

retention intervals typically have been used in examining t-trlal effectson motor task retention (Duffy, et al., 1975; Stelmach & Bassin, 1971).Because the effect of repeated t-trials on motor retention may vary withthe length of the interval as it does for verbal retention (Hogan &Kintsch, 1971; Thompson, et al., 1978, Exp 111), meaningful comparisonsbetween t-trial repetition effects on verbal and motor retention havebeen difficult to make.

The present experiment was designed to examine the relative effecti-veness of different motor training methods under acquisition and retentionconditions similar to those used in verbal training studies. The generalapproach was to allow either repetition or alternation of p- and t-trials prior to a given t-trial during training and to compare theeffects of such manipulation on both the acquisition and retention of alinear positioning movement.

In order to create acquisition and retention conditions similar tothose used in verbal task training, the present experiment differed fromother motor task training experiments in three ways. First, trainingprocedures were designed so that subjects were required to study thecriterion movement to be learned during p-trials and to recall it frommemory during t-trials. This was accomplished by using a constrainedmovement procedure at p--trial execution and a preselected movementprocedure at t-trial execution. Constrained p-trial movements wereperformed with the aid of a mechanical stop. Use of the stop ensuredthat subjects would study the criterion movement. Preselected t-trialmovements, on the other hand, were performed without the aid of themechanical stop. Removal of the stop ensured that subjects would haveto recall the to-be-learned movement from memory. Second, p- and t-trial training effects were measured over both short- and long-termretention intervals. This allowed for the examination of the potentialinteraction between training method and retention interval length andpermitted a more meaningful comparison of repeated t-trial trainingeffects on verbal and motor retention. Third, training was restrictedto the kinesthetic cue of distance. Although multiple cues underlie therecall of positioning movements (Hagman & Williams, 1977; Gundry, 1975),training was restricted to the specific cue of distance to prevent thepossibility of unsystematic subject selection of individual cues and thepossibility that certain cues might react differently to p- and t-trialrepetition because of their differential retention characteristics(Laabs, 1973; Posner, 1967).

METHOD

Subjects

Forty-five government employees (27 men and 18 women) volunteered toserve as subjects in the experiment. Forty-three were members of theprofessional and clerical staff of the Army Research Institute and twowere professionals affiliated with other agencies.

3

Apparatus

Movements were made from left to right using a metal slide whichslid along two stainless steel rods 35 in. (88.90 cm) in length. TwoThompson Ball Bushings supported the slide on the rods which were mountedin parallel on a metal frame 4.25 in. (11.00 cm) apart and located 11in. (27.94 cm) above the base of the frame. The base rested on astandard table top 31 in. (78.74 cm) from the floor. A second slide wasused to stop p-trial movements. This slide could be securely positionedby the experimenter along the entire length of the steel rods. Displace-ment of each slide rotated a separate 10-turn variable voltage precisionpotentiometer which was mounted on one side of the frame. The positionof each slide corresponded to a specific voltage level at the potentiometer

and was displayed by a digital voltmeter. The displayed voltage correspondedto millimeters and was accurate to within + 1 mm. Additional apparatusincluded a chin rest to control head movements and body position;earphones through which subjects heard tape-recorded procedural instructions;and a blindfold to prevent subjects from using visual cues during theexperiment.

Design

The experiment contained an acquisition and a retention segment asshown in Figure 1. The acquisition segment consisted of 18 trainingtrials divided into three cycles of six trials each. Each cycle containedp- and t-trials. P-trials were constrained movements during whichsubjects contacted a mechanical stop after moving the to-be-learneddistance of 250 mm. T-trials were preselected movements during whichsubjects attempted to recall the studied distance without the presenceof the mechanical stop. The sequence of p- and t-trials within cyclesdiffered for each of three training groups. For Group STANDARD, a cycleconsisted of three p- and three t-trials administered in an alternatingsequence. For Group PRESENTATION, the first five trials in each cyclewere p-trials and the sixth was a t-trial. For Group TEST, the firsttrial was a p-trial and the next five were t-trials. Training was suchthat a t-trial occurred every sixth trial for all three groups. As aresult, a 3x3 mixed factorial design was used to examine acquisitionperformance with the between-subjects variable being Groups (STANDARD,PRESENTATION, TEST) and the within-subjects variable being Trials (6,12, 18).

The retention segment of the experiment consisted of a single t-trial performed at both 3 min and 24 hr after Trial 18 of acquisition.Performance at Trial 18 was used to evaluate immediate recall accuracyand incorporated into a 3x3 mixed factorial design used to analyzeretention. In this design, the between-subjects variable was Groups(STANDARD, PRESENTATION, TEST) and the within-subjects variable was Time

4

of Recall (Immediate, 3 min, 24 hr). Fifteen subjects were assignedrandomly to each of the three groups with the constraint that each groupcontain the same number of men and women.

A total of 20 movements were performed during the experiment. Eachbegan from a different starting position which varied between 0 and 380mm from the left end of the apparatus in increments of 20 mm. Variationof movement starting position prevented the use of stopping location asan aid to learning distance. Three random starting position sequenceswere developed. Five subjects in each group were trained under one ofthe three sequences.

Procedure

At the beginning of the experiment, all subjects were instructed tolearn and remember distance. They were shown a written copy of theentire p- and t-trial command sequence appropriate to their traininggroup and told tile meaning of each command that they would he hearing.The p-trial command was "Movement" and the t-trial command was "RecallMovement." Each of these commands was preceded by "Ready" and followedby "Rest." At "Ready" the experimenter graspeQ the subject's right handand placed it on the handle of the slide. Five seconds later, subjectsheard either "Movement" or "Recall Movement" depending on their traininggroup. At the "Movement" command, they moved the slide across thelinear track at a moderate pace until contacting the mechanical stop.At the "Recall Movement" command, they moved the slide until they feltthat they had recalled the correct distance. Five seconds were allowedfor execution of p- and t-trial movements. During this intetval, whitenoise was delivered over the earphones to eliminate auditory cuLsresulting from displacement of the slide. "Rest" marked the start of a10 sec time interval during which subjects removed their hand.; from theslide and placed it in a predetermined resting place on the table. Itwas during rest periods that the experimenter recorded recall accuracyto the nearest mm and repositioned either the slide alone or both it andthe mechanical stop in preparation for the next trial. After "Rest"subjects heard "Ready" and the sequence of commandm for the next trialbegan. During the retention segment of the experiment, intervals of 3main and 24 hr were inserted between "Rest" and "Ready." Subjects wereasked not to count while moving the slide and were shown the approximatemovement speed (125 mm/sec) desired by the experimenter. Prior tomaking the first movement subjects donned their blindfold and earphoiic -;and were then given an opportunity to move the slide and get a feel forits basic movement characteristics.

RESULTS

Algebraic (signed) and absolute (unsigned) error scores were recordedfor each t-trlal performed during the acquisition and retentton segmentsof the experiment. The scores for each segment were aralyzefd separatelY.

5

.

CL- CL

LUCL

I-

>- CL CL.

U.' Co I- C- 0

cnn

CL . CL 0

q- 1 I-0

z q ~ g~ 0

.

LU CL

m% 0. . C0.0.

04.

0.

0I

CL.

C03 L0

Acquisition

Mean algebraic and absolute error scores for acquisition t-trialsare shown on the left in Figures 2 and 3. Although each figure depictsthe mean scores of all t-trials performed during acquisition, initialstatistical analyses were restricted to the scores on those t-trialswhich coincided temporally for all three training groups, i.e., trials6, 12, and 18. These scores were analyzed using a Groups (STANDARD,PRESENTATION, TEST) by Trials (6, 12, 18) mixed factorial analysis ofvariance.

Analysis of algebraic error revealed only a significant main effectof Groups, F (2,42) = 3.62, p < .05. Based on this significant F-value,individual comparisons were made using the least squared difference(LSD) method (Carmer & Swanson, 1973). All comparisons had a rejectionregion of .05. They revealed that TEST group performance was differentfrom that of the STANDARD, LSD (42) = 20.98, and PRESENTATION groups,LSD (42) = 19.80, but that the performance of these latter two groupsdid not differ. The STANDARD and PRESENTATION groups displayed asignificant overshooting bias with LSD (42) = 17.58 and 16.40, respectively,whereas the apparent undershooting tendency of the TEST group was notsignificant. Visual inspection of F'gure 2 reveals that algebraic errordifferences were already present between groups on the first t-trial ofthe first cycle and remained essentially unchanged throughout acquisition.Consequently, it could be argued that the obtained differences wereprobably a function of preexisting response bias differences presentamong the subjects in each training group rather than the result ofdifferential training. Apart from this apparent response bias differ-ence among groups, no other significant algebraic error effects werefound.

Analysis of absolute error revealed a significant main effect ofTrials, F (2,84) = 4.08, p < .05 and Groups, F (2,42) = 4.48, p < .05.Individual comparisons indicated that the Trials effect was causedprimarily by a decrease in absolute error between trials 6 and 12, LSD(84) = 8.11, as the error decrease between trials 12 and 18 was notsignificant. The Groups effect was the result of the TEST group havinggreater error than either the STANDARD, LSD (42) = 13.17 or PRESENTATIONgroups, LSD (42) = 14.29. The error for these latter two groups, however,did not differ. Thus, the groups which experienced either p-trialrepetition or p- and t-trial alternation during training performedbetter than the group which experienced t-trial repetition during training.Visual inspection of Figure 3 reveals that the curves of the STANDARDand PRESENTATION groups were similar to one another, while the curve ofthe TEST group was slightly above the other two and had a serratedappearance. This serrated curve of the TEST group reveals that absoluteerror tended to increase as t-trials were repeated within each acquisitiontrial cycle but that this increase was compensated for by a large decreasein error between cycles. To determine the reliability of the within-

7

ccJ

m.

Pu L.

z z

z£0 w

LL~uiuco- I.zL

CA CL

-

- C

M

a - cn

t- 76CI

cc E0

U0

03.

-i4-.

C.3C

C-

cp &n a in cm - "

(WWi) VOVUl3 31VUU391V NV3MI LL

P Cu

L0.

0 L.

LU

0

I--

LUU

0-

z I-

Cc Lu

z<I--

0n U)

~.0LU 0

P C-4 -j

C3 LU(aJ 5 C.)

.0

-D 0

L n .0

C.3 .. 0>- , -

C-4

C.) CT

CD

LU L

cycle error increase, a Cycle (1-3) by Trials (1-5) analysis of variancewas performed on all t-trial scores for the TEST group. As expected,the Trials effect was significant, F (4,56) = 4.24, p < .05. Individualcomparisons showed that absolute error increased reliably within cyclesand that performance on the last t-trial of each cycle was inferior tothat on the first t-trial of each cycle. This within-cycle error increasecaused by t-trial repetition is consistent with the results of previousmotor (Bilodeau & Bilodeau, 1958; Duffy, et al., 1975) and verballearning studies (e.g., Bregman & Wiener, 1970) which also have usedrepeated t-trials during training. To examine the between-cycle decreasein error, group performance was compared before within-cycle error had achance to materialize for the TEST group. This comparison required theuse of an additional Groups by Trials analysis of variance. In thisanalysis, STANDARD and PRESENTATION group performance at the end of eachcycle, i.e., trials 6, 12 and 18, was compared with TEST group performanceat the beginning of each cycle, i.e., trials 2, 8 and 14. Although thetrials effect was still significant with this analysis, F (2,84) = 4.78,p< .05, the Groups effect was not. The lack of a Groups effect indicatedthat TEST group error was greater than that of the other two groups onlyat the end of each cycle after within-cycle error had had a chance tobuild up. This build up of within-cycle error, however, was somehowcompensated for by a decrease in error between cycles. Thus, repeatingt-trials during training produced both positive and negative effects onacquisition. The negative effects took the form of increased within-cycle error and the positive effect took the form of decreased between-cycle error. Both of these effects have been reported previously byverbal researchers (e.g., Izawa, 1970; Tulving, 1967).

Retention

Mean algebraic and absolute error scores obtained during the retentionsegment of the experiment are shown on the right side of Figures 2 and3. Retention was examined using a Groups (STANDARD, PRESENTATION, TEST)by Retention Interval (Immediate, 3 min, 24 hr) analysis of variance.Performance at trial 18 was included in the analysis to indicate immediaterecall at the end of acquisition trials.

For algebraic error, no significant effects were found although theorder of group performance present at the end of acquisition was maintainedacross retention intervals. For absolute error, a significant Groups byRetention Interval interaction was found, F (4,84) - 4.10, R < .05. Asshown in Figure 3, this interaction resulted from an increase in erroracross retention intervals for the STANDARD and PRESENTATION groups anda decrease in error for the TEST group. Individual comparisons ofsimple effects revealed that at the end of acquisition (trial 18) theTEST group displayed greater error than either the STANDARD, LSD (84) =

19.00, or PRESENTATION group, LSD (84) = 17.86, and that no differenceexisted between the errors of these latter two groups (i.e., TEST >PRESENTATION = STANDARD). Three minutes after acquisition the only

10

significant change in group performance was an increase in error for thePRESENTATION group, LSD (84) = 16.07. As a result of this increase, theTEST and PRESENTATION group scores did not differ 3 min after acquisitionand their average absolute error was greater than that of the STANDARDgroup, LSD (84) = 13.83 (i.e., TEST = PRESENTATION > STANDARD). Boththe PRESENTATION and the STANDARD group displayed significant error increasesbetween 3 min and 24 hr after acquisition, with LSD (84) = 13.45 and13.40 for each group, respectively. In contrast, TEST group errordecreased significantly over this same time period with LSD (84) = 13.13. As aresult, 24 hr after acquisition TEST group performance was superior tothat of the STANDARD group, LSD (84) = 13.40 which, in turn, was superiorto that of the PRESENTATION group, LSD (84) = 15.80 (i.e., TEST < STANDARD <PRESENTATION). Thus, emphasis on p-trials during training resulted inrapid and extensive retention losses, whereas p- and t-trial alternationenhanced short-term retention while emphasis of t-trial repetitionenhanced long-term retention of movement distance.

DISCUSSION

The three training methods examined in the present research produceddifferent effects on motor acquisition and retention. During acquisition,performance improved in the usual negatively accelerated fashion whentraining consisted primarily of p-trial repetition or p- and t-trialalternation. However, when training consisted primarily of t-trialrepetition, performance was characterized by increased within-cycleerror offset by decreased between-cycle error. Because of the way thatperformance varied during acquisition when t-trials were repeated, therelative superiority of the three training methods was a function ofwhere their performance was compared during acquisition. When comparisonswere made before t-trial repetition had inflated the within-cycle errorof the TEST group, no performance differences were found among the threegroups. When comparisons were made after the TEST group had completedt-trial repetition, the training methods employing p-trial repetitionand p- and t-trial alternation produced superior acquisition performance.

Of particular interest is the question of why t-trial repetition -produced both increased within-cycle error and decreased between-cycleerror during acquisition. There are at least two reasons for theincreased error within cycles. First, subjects may have been attemptingto recall an ever-decaying memorial representation of movement distanceestablished at the execution of an earlier p-trial, and as a result,their recall accuracy got progressively worse. This notion is consistentwith short-term motor memory research showing that distance informationdoes decay rapidly and cannot be retained over even short retentionintervals (e.g., Laabs, 1973). Second, within-cycle error could haveincreased because t-trial repetition produced interference (Bilodeau &Bilodeau, 1958). Although t-trials are not intended to interfere withretention, they usually are different from the criterion movement, andtherefore possess the potential for producing interference (Hlagman,

ii

k .

1978). Assuming a positive relationship between the magnitude ofabsolute error and the number of interfering movements performed, increasedwithin-cycle error could have been a function of the added number ofinterfering movements afforded by t-trial repetition. Because bothdecay and interference have adversely affected retention of distance inthe past, it is likely that both processes contributed to the within-cycle error increases observed in the present experiment. One way toexamine the effect of interference alone would be to perform the sameexperiment using a movement cue such as stopping location which isaffected adversely by interference (Hagman, 1978) but not by decay(Laabs, 1973).

Reasons why t-trial repetition produced such large between-cycledecreases in error are not readily apparent. Researchers in verballearning have suggested that one function of testing via recall is toenhance a subject's ability to recognize past words recalled (Klee &Gardiner, 1976). Perhaps, the same effect of testing occurs in motorlearning. After repeated testing attempts, subjects may know more abouttheir recall performance than those subjects who do not perform repeatedtesting. Consequently, subjects in the TEST group may have been betterable to discriminate their recalled distance from that of the criteriondistance and been more capable of making the appropriate adjustmentsneeded for more accurate recall. Although speculative, the notion thatincreased discrimination ability results from t-trial repetition isconsistent with previous motor research (Newell, 1974). In essence, t-trial repetition may serve to potentiate the effect of subsequent p-trials. This type of potentiating effect has been reported to occur inverbal learning (Izawa, 1970) but has only been hinted at in motorlearning (Bilodeau & Bilodeau, 1958; Henderson, 1977). Thus, additionalresearch is needed before a firm conclusion can be drawn about thepotentiating effects of t-trial repetition in motor learning. Oneapproach might be to vary the number of t-trials repeated prior to p-trial execution. If potentiation does occur, one should find that asthe number of repeated t-trials increases performance following asubsequent p-trial should also increase.

To better examine potentiation effects in motor learning, a taskmore difficult than linear positioning should be chosen. This wouldreduce the possibility of unwanted ceiling or floor effects influencingthe data. In the present experiment, acquisition performance of theSTANDARD and PRESENTATION groups tended to flatten out both within andbetween cycles 2 and 3. Because of this, it was possible that ceilingeffects on accuracy prevented future reduction of errors for these twogroups. TEST group performance, however, was not near the ceiling atthe end of each cycle due to within-cycle error increases, and thus, hadample room for improvement between cycles. By eliminating any possibilityof ceiling effects differentially affecting group peiformance, potentiationcould be examined more effectively.

12

The different influence of p- and t-trial training methods was mostevident at retention. Forgetting of distance was rapid and extensivewhen p-trials were repeated during training. Repetition of t-trialsduring training, however, retarded long-term forgetting and actuallyproduced improvements in recall accuracy over time. Alternation of p-and t-trials during training prevented short-term retention losses butwas not as effective a3 t-trial repetition in reducing long-term retentionlosses. Apparently, a greater emphasis on t-trial repetition duringtraining is needed to prevent long-term forgetting of movement distance.

Two questions regarding the retention results need to be discussed.First, why did repeated t-trial training at acquisition cause superiorlong-term retention of movement distance? And second, why did -epeatedt-trial training cause improved retention over time? Although it wouldbe premature to suggest any definite answers to these questions beforeadditional research is conducted, some mention of a potential explanatoryconcept for each question would seem appropriate at this time.

Consider the first question of why did t-trial repetition producesuperior long-term retention of distance. A potential answer to thisquestion relies on the distinction between constrained and preselectedmovement procedures and the types of information available to subjectsfor processing as a result of executing constrained p-trials and pre-selected t-trials. Investigators have shown that constrained movementsprovide subjects with kinesthetic feedback information originating frommuscles and joints (Goodwin, McCloskey & Matthews, 1972; Marteniuk &Roy, 1972). Preselected movements, on the other hand, not only providethis kinesthetic feedback but also provide additional information regardingefferent commands given to the muscles (Jones, 1974). Presumably, foraccurate efferent-command information subjects must know the movementstopping location in advance (Stelmach, Kelso, & McCullagh, 1976).Hence, in the present experiment accurate efferent information waspresent only under preselected movement procedures. The presence ofefferent information allows subjects to preprogram their recall ofdistance (,Jones & Hulme, 1976) and because of this distance is retainedbetter when performed under a preselected rather than a constrainedmovement procedure (Kelso, 1977; Stelmach, Kelso & Wallace, 1975).Because of the superior retention characteristics associated with pre-selected movements, subjects may have relied on retention of theirrecall performance at t-trials rather than their study performance at p-trials to support later t-trial recall attempts. If long-term retentionof t-trial recall performance improves as a function of t-trial repetitionas might be expected, then the order of group recall error 24 hr afteracquisition would have been as reported (i.e., TEST < STANDARD < PRESEN-TATION).

One problem with this interpretation is the relatively low errorlevel shown by the TEST group 24 hr after acquisition. If retention ofrecall performance were a function of prior t-trial repetition, one

13

V

might expect a performance error level similar to either the averageerror level of all prior t-trials or to the error level of the last t-trial of acquisition. Instead, the error level was similar to that ofthe first t-trial of the last acquisition trial cycle (trial 14). Itappears as though TEST group subjects tried to duplicate the recall

performance of acquisition trial 14. If so, this would explain theirrelatively low recall error level. To support this notion, otherresearchers (Bilodeau, Jones & Levy, 1964) have found that when a seriesof t-trials follows a p-trial (as in Cycle 3 of acquisition), subjects

tend to recall their performance on the first t-trial of the seriesrather than their performance on either the p-trial or other adjacent t-

trials. It is as though t-trial repetition within-cycles produces aform of serial learning task where subjects try to remember their recall

performance at each successive trial. Because of the strong primacyeffects usually associated with serial motor tasks (Magill & Dowell,1977), subjects are better able to remember their recall performance on Vthe first t-trial of the series and rely on it for later recall.Additional support for the notion that knowledge of recall performance4

is better for the initial items of a serial task has been reported forverbal task learning (Klee & Gardiner, 1976; Lockhart, 1975).

A possible answer to the second question of why did TEST grouprecall performance improve across retention intervals involves thenotion of retroactive interference. If the assumption is correct that

TEST group subjects did attempt to duplicate their recall performance ofacquisition trial 14, then the decrease in absolute error experiencedbetween 3 min and 24 hr after acquisition was caused by the dissipationof retroactive interference. This retroactive interference was generatedinitially by added acquisition t-trials performed after trial 14 and wasof sufficient strength to degrade recall performance both at the end of

acquisition and at the 3 min retention interval. However, its strengthapparently dissipated thereafter and did not adversely affect retention24 hr after acquisition. Although dissipation of retroactive interferencehas primarily been found in verbal retention studies (e.g., Postman,Stack & Fraser, 1968), the general viewpoint that the strength of inter-item interference does change over time has received support in recentmotor retention studies (e.g., Wrisberg, 1975).

CONCLUSION

In conclusion, the results of the present experiment clarify certaintraining issues regarding the relative effects of presentation andtesting on motor task acquisition and retention. In doing so, they

answer the Army's question of which specific training methods mosteffectively promote the highest levels of skill acquisition and retention.

The results indicate that, first, training methods which eitherrepeat or alternate presentation and testing during acquisition enhancemotor acquisition performance. Final acquisition performance is better,

14

W.L

however, when training is conducted using methods which emphasize eitherrepeated presentation or alternation of presentation and testing of theto-be-learned movement rather than its repeated testing.

Second, the primary benefit of repeating testing during training isenhanced long-term retention. Relative to the other two training methods,long-term motor retention is improved substantially by a training methodwhich allows opportunities for repeated testing during acquisition.Thus, the viewpoint that regards testing as merely an opportunity toshow what has been learned during presentation is incorrect. Apparently,there are benefits associated with repeated testing which are derivedfrom the additional memorial information generated by the process ofactive recall. Recall serves to modify memory such that prior recall ofmemorized motor information facilitates later recall of this same infor-mation. This facilitation is greater than that resulting from repeatedstudy of motor information during presentation.

Third, training methods which manipulate presentation and testingopportunities during acquisition have a similar effect on the acquisitionand retention of both motor and verbal tasks. This is true, at least,when presentation and test procedures employed during training aresimilar for both types of tasks.

Fourth, the benefits of repeated testing have been 3hown for thestandard laboratory motor task of linear positioning. Future researchshould be directed toward answering the question of whether or not thesebenefits will generalize to other types of motor tasks. Of particularinterest should be the examination of procedural motor tasks whichrequire the execution of successive simple motor movements in the correctserial fashion and are characteristic of many motor tasks performedwithin the Army.

Fifth, and lastly, testing should be viewed as an effective way ofenhancing the long-term retention of motor information. This enhancedretention can be achieved by changing the emphasis of training frompresentation to testing and would occur without the negative aspects ofadditional expenditures in training time. money and personnel.

15

REFERENCES

Adams, J. A., & Dijkstra, S. Short-term memory for rotor responses.Journal of Experimental Psychology, 1966, 71, 314-318.

Allen, G. A., Mahler, W. A., & Estes, W. K. Effects of recall testson long-term retention of paired associates. Journal of VerbalLearning and Verbal Behavior, 1969, 8, 463-471.

Bilodeau, E. A., & Bilodeau, I. Mc D. Variable frequency of knowledgeof results and the learning of a simple skill. Journal ofExperimenter Psychology, 1958, 55, 379-383.

Bilodeau, E. A., Jones, M. B., & Levy, C. M. Long-term memory as afunction of retention time and repeated recalling. Journal ofExperimental Psychology, 1964, 67, 303-309.

Bjork, R. A. Retrieval as a memory modifier: An interpretation ofnegative recency and related phenomena. In R. L. Solso (Ed.),Information processing and cognition the Loyola symposium, HillsdaleNew Jersey: Lawrence Erlbaum Associates, Publishers, 1975.

Bregman, A. S., & Wiener, J. R. Effects of test trials in paired-associate and free-recall learning. Journal of Verbal Learningand Verbal Behavior, 1970, 9, 689-698.

Carmer, S. G., & Swanson, M. R. An evaluation of ten pair-wise multiplecomparison procedures by Monte Carlo methods. Journal of theAmerican Statistical Association, 1973, 68, 66-74.

Dosher, B. A. & Russo, J. E. Memory for internally generated stimuli.Journal of Experimental Psychology: Human Learning and Memory,1976, 2, 633-640.

Duffy, T. M., Montague, W. E., & Laabs, G. J. The effect of overtrehearsal on motor short-term memory. Journal of Motor Behavior,1975, 7, 59-63.

Gates, A. I. Recitation as a factor in memorizing. Archives ofPsychology, 1917, 40, 104.

Goodwin, G. M., McCloskey, D. I., & Matthews, P. B. C. The contributionof muscle afferents to kinesthesia shown by vibration induced illusionsof movement and by the effects of paralyzing joint afferents.Brain, 1972, 95, 705-748.

Gundry, J. The use of location and distance in reproducing differentamplitudes of movement. Journal of Motor Behavior, 1975, 7,91-100.

17 IACG pI & AN ( LOT i.

Hagman, J. Specific-cue effects of interpolated movements on distanceand location retention in short-term motor memory. Memory andCognition, 1978, 6, 432-437.

Hagman, J. D., & Williams, E. Use of distance and location in short-term motor memory. Perceptual and Motor Skills, 1977, 44, 867-873.

Hellyer, S. Supplementary report: Frequency of stimulus presentationand short-term decrement in recall. Journal of ExperimentalPsychology, 1962, 64, 650.

Henderson, S. E. Role of feedback in the development and maintenanceof a complex skill. Journal of Experimental Psychology: HumanPerception and Performance, 1977, 3, 224-233.

Hogan, R. M. & Kintsch, W. Differential effects of study and testtrials on long-term recognition and recall. Journal of VerbalLearning and Verbal Behavior, 1971, 10, 562-567.

Holding, D. H., & Macrae, A. W. Guidance, restriction and knowledge ofresults. Ergonomics, 1964, 7, 289-295.

Izawa, C. Optimal potentiating effects and forgetting-prevention effectsof tests in paired-associate learning. Journal of ExperimentalPsychology, 1970, 83, 340-344.

Jones, B. Role of central monitoring of efference in short-term memoryfor movements. Journal of Experimental Psychology, 1974, 102,

37-43.

Jones, B., & Hulme, M. R. Evidence for an outflow theory of skill.Acta Psychologica, 1976, 40, 49-56.

Kelso, J. A. S. Motor control mechanisms underlying human movementreproduction. Journal of Experimental Psychology: Human Perceptionand Performance, 1977, 3, 529-543.

Klee, I., & ..ardlner, J. N. Memory for remembered events: Contrastingrecall -d recognition. Journal of Verbal Learning and Verbal

Behavior, 1976, 15, 471-478.

Laabs, G. J. Retention characteristics of different reproduction cuesin motor short-term memory. Journal of Experimental Psychology,1973, 100, 168-177.

Lockhart, R. S. The facilitation of recognition by recall. Journalof Verbal Learning and Verbal Behavior, 1975, 14, 253-258.

Magill, R. A., & Dowell, M. N. Serial-position effects in motor short-term memory. Journal of Motor Behavior, 1977, 9, 319-323.

18

Marteniuk, R. C., & Roy, E. A. The codability of kinesthetic locationand distance information. Acta Psychologica, 1972, 36, 471-479.

Newell, K. M. Knowledge of results and motor learning. Journal ofMotor Behavior, 1974, 6, 235-244.

Newell, K. M. Motor learning without knowledge of results through thedevelopment of a response recognition mechanism. Journal ofMotor Behavior, 1976, 8, 209-217.

Posner, M. I. Characteristics of visual and kinesthetic memory codes.Journal of Experimental Psychology, 1967, 75, 103-107.

Postman, L., Stark, K., & Fraser, J. Temporal changes in interference.Journal of Verbal Learning and Verbal Behavior, 1968, 7, 692-694.

Raffel, G. The effect of recall on forgetting. Journal of ExperimentalPsychology, 1934, 17, 828-838.

Rosner, S. R. The effects of presentation and recall trials on organizationin multi-trial free recall. Journal of Verbal Learning and VerbalBehavior, 1970, 9, 69-74.

Seashore, R. H., & Bevelas, A. The functioning of knowledge of resultsin Thorndikes line drawing experiment. Psychological Review,1941, 48, 155-164.

Stelmach, G. E., & Bassin, S. L. The role of overt motor rehearsal inkinesthetic recall. Acta Psychologica, 1971, 35, 56-63.

Stelmach, G. E., Kelso, J. A. S., & McCullagh, P. D. Preselection andresponse biasing in short-term motor memory, Memory and Cognition,1976, 4, 62-66.

Stelmach, G. E., Kelso, J. A. S., & Wallace, S. A. Preselection inshort term memory. Journal of Experimental Psychology: HumanLearning and Memory, 1975, 1, 745-755.

Thompson, C. P., Wenger, S. K., & Bartling, C. A. How recall facilitatessubsequent recall: A reappraisal. Journal of Experimental Psychology:Human Learning and Memory, 1978, 4, 210-221.

Thorndike, E. L. The law of effect. American Journal of Psychology,1927, 39, 212-222.

Tulving, E. The effects of presentation and recall of material infree recall learning. Journal of Verbal Learning and VerbalBehavior, 1967, 6, 175-184.

19

Wrisberg, C. A. The serial-position effect in short-term motor

retention. Journal of Motor Behavior, 1975, 7, 289-295.

Wrisberg, C. A., & Schmidt, R. A. A note on motor learning without

post-response knowledge of results. Journal of Motor Behavior,1975, 7, 221-225.

20

L

DISTRIBUTION

ARt Distribution List

4 QASO (M&RAI 2 HOUSACOEC, Ft Ord, ATTN: Library2 IIOOA IDAMI.CSZ) 1 IIOUSACDEC. FP Ord, ATTN: ATEC-EX-E-Hum Factors1 HODA IOAPE-PSNI 2 USAEEC, Ft Beonjamin Harrison, ATTN: LibraryI HODA (DAMA.AR) 1 USAPACDC. Ft Benjamin Harrison. ATTN: ATCP-HR1 HODA (DAPE-HRE-POI 1 USA Comm-Elect Seft, Ft Monmouth, ATTN: ATSN-EAI HODA (SGODID) 1 USAEC. Ft Monmouth, ATTN: AMSEL-CT-HDPI HODA (OAMI-DOT-C) 1 USAEC. Ft Monmnouthi, ATTN: AMSEL-PA-P1 HOA IDAPC.PMZA) 1 USAEC. Ft Monmouth. ATTN: AMSEL-St-CB1 HODA IDACH-PPZ.Al I USAEC. Ft Monmouth, ATTN: C, Fad Dev Br1 HOOA (DAPENHRE) 1 USA Materials Sys Anal Agcv, Aberdeen. ATTN: AA4XSY-P1 HC2DA IDAPE MPO.CI 1 Edgawood Arsenal. Aberdeen. ATTN: SAREA-BL-H1 HODA (DAPE-OWl I USA Ordi Cit & Scti. Aberdeen. ATTN: ATSL-TEM-C1 HOOA IDAPE-HRL 2 USA Hum Engr Lab. Aberdeen, ATTN: Library/Dir1 HODA (OAPE-CPS) 1 USA Combat Arma Trig Sd. Ft Banning. ATTN: Ad Supervisor1 HOOA (DAFO.MFA) 1 USA Infantry Hum Rsch Unit, Ft Banning. ATTN: Chief1 HODA (DARD-ARS.P) 1 USA Infantry Sd. Ft Banning, ATTN: STEBC-TE-TI HODA (DAPC-PAS-A) 1 USASMA. Ft Bliss. ATTN: ATSS-LRCI HODA (DUSA-OR) t USA Air Dof Schs, Ft Bliss. ATTN: ATSA-CTD-ME1 HODA (DAMO.ROR) 1 USA Air Def Sch. Ft Bliss. ATTN: Tech LibI HODA IDASG) I USA Air Delf 8d. Ft Bliss. ATTN: FILES1 HOA (DAIa-Pt) 1 USA Air Def Sd. Ft Bliss. ATTN: STEUD-POI Chief. Consult Div IDA-OTSG). Adelphi. MD 1 USA Cntd So General StI Collage. Ft Leavenworth, ATTN: LibI Mil Asnt. Hum Res. OODR&E. DAD (EdiLSI 1 USA Cmd & General Stf Collage. Ft Leavenworth. ATT'N: ATSW-SE-L1 HO USARAL, APO Seattle. ATTN: ARAGP-R 1 USA Cnsd So General Stf Collage. Ft Leavenworth,. ATT'N: Ed AdvisorI HO First Amy. ATTN, AFKA-OI.Tt I USA Combined Arms Cmt Dev Act. Ft Leavenworth. ATTN: DoapCdr2 HO Fifth Ai -v, Ft Sun Houston I USA Combined Arms Cmbt Dev Act. Ft Leavenworth. ATTN: CCS1 Dir. Army Stf Studies Ofc. ATTN: OAVCSA (DSP) I USA Combined Arms Cmbt Dert Act. Ft Leavenworth. ATTN: ATCASAI Oft: Chief of StI. Studies Ofc I USA Combined Arms Cmbt Dew Act, Ft Leavenworth. ATTN: ATCACO-E1 DCSPER. ATTN: CR5/OCR I USA Combined Arms Cmnbt Day Act. Ft Leavenworth. ATTN: ATCACC-CII The Army Lib. Pentagon, ATTN: ASS Chief 1 USAECOM, Night Viion Lab. Ft 8alvoir. ATTN: AMSEL-NV-SDI The Army Lib. Pentagon. ATTN: ANAAL 3 USA Compoute Sys Cmd, Ft Balvoir. ATTN: Tech Library1 Ofc. Ant Sa of the Army (R&DI 1 USAMEROC. Ft Balvoir. ATTN: STSF3-DO1 Tech Support 01c. OJCS I USA Eng Sch. Ft Balvoir, ATTN: Library1 USASA, Arlington. ATTN: IARO.T I USA Topographic Lab. Ft Beolsoir. ATTN: ETL-rO-SI USA Rich Olc, Durham. ATTN: Life Sciences Dir I USA Topographic Lab. Ft Balvoir. ATTN: STINFO Canter2 USARIEM. Natick,. ATTN: SGRO-UE*CA 1 USA Topographic Lab. Ft Beilvoir. ATTN: ETL-GSL1 USATTC. Ft Clayton, ATTN: STETC-MO-A 1 USA Intelligence Cu, & Sch. Ft Huactiuca. ATTN: CTO-MS1 USAIMA. Ft Bragg. ATTN: ATSU.CTO-OM I USA Intelligence Ctr &i Sch. Ft Huachuca. ATTN: ATS--C T-MS1 USAIMA. Ft Bragg. ATTN: Marquat Lib 1 UISA Intelligenc Ctr &i Sch. Ft Huachuca. ATTN: ATSI-TEI US WAC Ctr &i Sch. Ft McClellan. ATTN: Lib I LISA Intelligence Ct Ai Sdh. Ft Huechucs. ATTN: ATSI-TFX-OSI US WAC Ctr & Sch. Ft McClellan. ATTN: Tnig Dir I USA Intelligence Ctr Go Sch. Ft Huachuca. ATTN! ATSI-CTS-OR1 USA Ouarteriaster Sch. Ft Lee, ATTN: ATSMI-TE 1 USA intelligence Ct So Sch. Ft Huaichuca., ATTN: ATSI-CTD-DT1 intelligence Material Dee Ofc. EWL. Ft Holabird I USA intelligence Ctr & Sch. Ft Huachuca. ATTN: ATSI-CTO-'CS1 USA SE Signal Sch. Ft Gordon. ATTN: ATSO.EA I USA Intelligence Ctf So Sch, Ft Huaciuca, ATTN: DAS/SAD1 USA Chaplain Csr & Sd'. Ft Hamilton. ATTN: ATSC-TE-RD 1 USA Intelligence Ct &i Sch. Ft Nuachuca., ATTN: ATSI-TEM1 USATSCH. Ft Eustis, ATTN: Educ Advisor I USA Intelligence Ct So Sch. Ft Huachsjce. ATTN: LibraryiUSA War Collage. Carlisle Barracks. ATTN: Lib I COR, HO Ft Huachuca. ATTN: rowc Relf Div

2 WRAIR, Neuropsyrchiatry Div 2 COR. UISA Electronic Pr"g Grd. ATTN: STEER-MT-SI DLI. SOA. Monterey I HO. TCATA. ATTN: Tecs LibraryI USA Concept Anal Agcy. Bethesda, ATTN: MOCA-MR I HO. TCATA. ATN: AT CAT-OP-a. Ft Hood1 USA Concept Anal Agcy. Bethesida. ATTN: MOCA..JF I USA Recruiting Cmd. Ft Sheridan. ATTN: USARCPPI USA Arctic Teat Ctr, APO Seattle. ATTN: STEAC-RL-MI I Senior Army Ada.. USAFAGOD/TAC. Elgot AF Aux FId No. 91 USA Arctic Test Ctr. APO Seattle, ATT'N: AMSTE-PL-TS I HC. USARPAC. DCSPERA. APO SF 966560. ATTN: GPPE -SE1 USA Armament Cnsd, Redstone Anrseal. ATTN: ATSK.TEM I Stimaco Lib. Academy of Health Sciences . Ft Sam HouatonI USA Armament Cmd. Rods Island. ATTN: AMSAR-TDC I Marine Corps Inst.. ATTN: Den-MCII FAA.NAFEC, Atlantic City. ATTN: Library I HQ. USMIC. Commandant. ATTN: Code MTMTI FAA.NAFEC. Atlantic City. ATTN: Human Engr Br 1 HO. USMC, Commandant. ATTN! Code MPI-20-28I FAA Aeronautical Ctr. Oklahoma City. ATTN: AAC-440 2 USCG Academy. New London. ATTN: Admisaion2 USA Fid Arny Sch. Ft Sill. ATTN: Library 2 USCG Academy. New Lonion. ATTN: LibraryI USA Armor Sch. Ft Knox. AMN: Library I USCG Training Cir. NY. ATTN: COI USA Armor Sd'. Ft Knox. ATTN: ATSB-01-E I USCG Training Ctr, NY, ATTN: Educ; Svc OfcI USA Armor Schi, Ft Knox. ATTN: ATSB.OT-TP 1 USCG. Psythol RA" Sr. DC. ATTN: GP 11621 USA Armor Sch. Ft Knox. ATTN: ATSS.CO*AD I HO Mid-Range Br. MC DOn. Ousansico. ATTN: P&S Div

21

I US Marine Corps Liaison Ofc. AMC. Alexandria. ATTN: AMCGS-F I Otf Civil nit of Enviro Medicine. CanadaI USATRADOC. Ft Monroe. ATTN! ATRO-ED I AIR CRESS. Keansinlgton, ATTN -info Sys Sr

6 USATRAOOC. Ft Monroe. ATTN: A TPA -AD I Militastroiilogisi Tionesta CopenhagenI USATRAQOC, Ft Monroe. ATTN! ATTS-EA I M~iitary Attache. French Embassy. ATTN' Ooc SeeI USA Form~ Cmil. Ft McPhserson, ATTtc Libri I Meidici's Chat C.ER P.A-Arsenal. TosulonrNaval France

2 USA Aviation Test Sd. Ft Rfucker, ATTN STEEG-PO I Prin Scientitic Off. Apod Hum Engr Rscy Div. Ministry

I USA Agcy for Aviation Safety. Ft Rucker, ATTN: Library of Defense. New Delfhi

I USA Agicy for Aviation Safety. Ft Ruckxer. ATTN: Educ Advisor 1 Per$ Rsch Ofc Library. AKA. Israel Defense Forces

1 USA Aviation Scfs, Pt Ruckxer. ATTN: P0 Draer 0 1 Ministeris van Deifenie. OOOPJKL Afo Social

1 HfOUSA Aviation Sys Cmd., St Louis. ATTN: AMSAV-ZDR Psychologische Zaken. The Hague, Netherlands

2 USA Aviation Svs Test Act.. Edwards AFG. ATTN: SAVTE-TI USA Air Dot Salt. Pt Bliss. ATTN: ATSA TEMI USA Air Mobility Rich & Des Lab. Moffett Fid. ATTN: SAVDL-ASI USA Aviation Sds. Res Trg Mgt, Ft Ruckxer, ATTN: ATST-T-RTMI USA Aviation Son. CO. Ft Rucker. ATTN: ATST-O-A1 HO. DARCOM. Alexandria. ATTN AMXCD-TLI HO. DARCOM. Alexandria. ATTIN. CDP1 US Military Academy. West Point. ATTN: Serials Unitt US Military Academy. West Point. ATTN: Ofc of Milt LdrshO1 US Military Academy. West Point. ATTN: MAOR1 USA Standardization Gp. UK. FPO NY. ATTN: MASE-GCI Qfc of Naval Rich. Arlington. ATTN : Code 4S23 Ofc of Navel Rich., Arlington. ATTN: Code 4581 Ofic of Naval Rich. Arfington. ATTN: Coda 4501 Otc of Naval Rsch, Arlington. ATTN: Code 4411 Naval Aetrospc Mod Res Lab. Pensacola, ATTN . Acous Sch Div1 Naval Aatrosoc Mad Res Lab. Pensacola, ATTN: Code L511 Naval Aerosoc Mad Ret Lao. Pensacola. ATTN: Code LBI Chief of NavPer, ATTN: Port-ORI NAVAIRSTA. Norfolk. ATTN, Safety Ctr1 Nay Oceanographic, DC. ATTN: Code 6251. Charts & TechI Center of Naval Anal, ATTN: Doc Ctr1 NavAirSysCom. ATTN: AIR-5313CI Nav BuMed, ArTN: 7131 NavHelicooterSubSqua 2. FF0 SIF 946011 AFHRL PFT) Williams AFBI AFHRL (TT) Lowry APSI AFHRL IASI WPAFB. OH2 AFHRL (DOjZ) Brooks AFB1 AFHRL IDOJNI Ladlxland AFSI HOUSAF (INYSOII HOUSAF (OPXXA1I AFVTG iRDI Randolph AFB3 AMRL iHEI WPAFB. OH2 AF Inst of Tech. WPAF8, OH. ATTN: ENEiSLI ATC IXPTDI Randolph AF81 USAF AeroMed Lb. Brooks AFB ISUL-4l. ATTN: DOC SEC1 AFOSR INL). Arlington1 AF Log Cmd. McClellan AFS. ATTN. ALC.'DPCRBI Air Force Academy, CO. ATTN- Dept of Bel Scn5 Nawers & Dev Cth. San Diego2 Navy Moo Neuropsyctriatric Rich Unit. Sani Diego

1Nav Electronic Lab. San Diego. ATTN: Res LabI Nay TrngCen. San Diego, ATrTN Code 9000-LibI Navfttlra~ch. Monterey. ATTN: Coo 5bAsI NaviststGralich. Monitrey. ATTN. Code 2124I NavrrnqEsuioCtr. Orlando. ATTNI: Tech LibI US Dept of Labor. DC. ATTN: Manpower Admin1 US Dept of Justice. DC. ATTN: Drug Enforce Admin1 Nat Bur of Standards. DC. ATTN: Computer Info Section1 Nas Clawing House for MH-linfo. RockrvilleIDenver Federal Ctr. Lakeawood. ATTN BLUM

2 Cefense Documentation Canter41 Dir PwCrr Army, HO. Russel Ofcs. Canbterry

Scenvtic Adyir. Mil Sd Armv Hn. Russeil Ofcs. Canoerras4.. ii'o Air Attache, Austrian Emrssv

Cntre rie Rechierche 0e; Facrurs. Huinaine de 'a Defense'4avor,aie 3russeis

2 Canadian Joint Staff WasingtonI .' Ar Staff Royal Canadian AF ATTN Pers Std Anal 8,

3 Cheof C.anadian Ott Rscn Staff ATTN- C CROSIWII1 9r-r.m 0*i Staff B' ttin E~ioasiv Wasngton

22