UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The development of a single-item test as a measure of soccer skill Johnson, Joseph Robert 1963

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1963_A7_5 J6 D3.pdf [ 3.47MB ]
Metadata
JSON: 831-1.0077413.json
JSON-LD: 831-1.0077413-ld.json
RDF/XML (Pretty): 831-1.0077413-rdf.xml
RDF/JSON: 831-1.0077413-rdf.json
Turtle: 831-1.0077413-turtle.txt
N-Triples: 831-1.0077413-rdf-ntriples.txt
Original Record: 831-1.0077413-source.json
Full Text
831-1.0077413-fulltext.txt
Citation
831-1.0077413.ris

Full Text

THE DEVELOPMENT OP A SINGLE-ITEM TEST AS A MEASURE OF SOCCER SKILL by JOSEPH ROBERT JOHNSON B.P.E. University of British Columbia, 1962. A THESIS SUBMITTED IN PARTIAL FULFILMENT OP THE REQUIREMENTS FOR THE DEGREE OP MASTER OF PHYSICAL EDUCATION We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1963. In presenting this thesis i n p a r t i a l fulfilment of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library sh a l l make i t freely available for reference and study. I further agree that per-mission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It i s understood that copying or publi-cation of this thesis for f i n a n c i a l gain shall not be allowed without my written permission. Department of The University of B r i t i s h Columbia, Vancouver 8, Canada. Date ABSTRACT The purpose of this study was to develop and establish a wall-volley type test as a method of measuring soccer s k i l l . The subjects were students in attendance at The University of British Columbia. The test was administered to 75 students who represented five distinct soccer ability groups: The Thunderbirds (Varsity f i r s t team), the Chiefs (Varsity second team), the Braves (Varsity third team), a Physical Education major class, and a Service Programme class. Each group consisted of 15 subjects. Subjects were rank-ordered by the experimenter according to ability. The test required each subject to perform the wall volley test of three 30-second t r i a l s . The subject's aggregate score was correlated against the experimenter's rank ordering of players. This test differed from previous wall-volley type tests of soccer ability i n the dimensions of the target area; the distance of the restraining line; the use of a moving ball at the commencement of the test, and in the method of scoring. Test results proved satisfactory, and the test suggests it s e l f as a speedy, economic means of evaluating the soccer s k i l l of players by coaches and physical educators. It was noted from repeated testing that subjects perform better after at least two practice sessions. TABLE OF CONTENTS CHAPTER PAGE I STATEMENT OF PROBLEM 1 II JUSTIFICATION OF PROBLEM 4 III REVIEW' OF LITERATURE 6 IV METHODS AND PROCEDURES 16 V PRESENTATION OF DATA 28 VI ANALYSIS OF DATA 34 VII SUMMARY, CONCLUSIONS AND RECOMMENDATIONS . . . . 43 APPENDICES I TEAM PERFORMANCE SCORES 49 II STATISTICAL TREATMENT 55 III GRAPH 64a BIBLIOGRAPHY 66 CHAPTER 1 STATEMENT OF PROBLEM Physical educators have long realized the value of s k i l l tests as a means of more objective measurement, and they have also recognized the d e s i r a b i l i t y of employing s c i e n t i f i c and sound procedures to obtain better results i n t h e i r work. However, although the area of tests and measurement i s given l i p service as being essential i n the evaluation of physical education programmes, many physical educators r e s i s t the use of tests on the grounds of u n r e l i a b i l i t y or non-validity. Many teachers also f e e l that the administration of tests i s too time consuming and requires too much supervisory control. In 1933, one of North America's leading physical educators, J . B. Nash, ( l ) wrote "the game of soccer i s rapidly gaining acceptance across North America and has already replaced f o o t b a l l i n many of our schools." This statement reveals the popularity of soccer t h i r t y years ago. Today the game enjoys considerable acceptance i n a l l forms of educational i n s t i t u t i o n s : the elementary school, junior and senior high schools, and colleges and u n i v e r s i t i e s . In Canada, as a result of the Duke of Edinburgh's address to the Canadian Medical Association Convention i n 1959, and more recently, the introduction by the Federal Government of B i l l C-131 (2) directed towards the promotion of National Fitness and Amateur Sport, greater interest has been aroused i n programmes of physical education and subsequently i n the objective evaluation of such programmes. 2 The problem here was to provide a quick, r e l i a b l e and economic method of testing the soccer s k i l l of University students. There was also the necessity of providing a test which would serve the layman soccer coach as well as the physical education teacher* I t was hypothesized then, that a wall volley-type test can be developed which w i l l e f f e c t i v e l y measure soccer s k i l l , and provide an objective means of c l a s s i f y i n g players. Statement of Problem: The problem i s to develop a te s t of general b a l l control a b i l i t y using as nearly as possible actual game s k i l l s . The purpose of th i s test i s to measure the soccer s k i l l of the individual and to serve as a means for c l a s s i f y i n g players. Sub-Problems Were: (a) To decide the dimensions of the target area. (b) To decide the distance of the restraining l i n e from the target area. (c) To determine the method of administering the t e s t . (d) To develop., an interim scoring scale as a measure of the soccer s k i l l of players. (e) To determine the effects of practice or learning by repeated re-testing. 3 REFERENCES 1. Caswell, John, E., "Preface", Soccer for High Schools, A. S. Barnes & Co., New York, 1933. 2. B i l l C-131, An Act to encourage Fitness and Amateur Sports. As passed by the House of Commons, 25th September, 1961. Queen's Printers, Ottawa, 1961. CHAPTER II JUSTIFICATION OF PROBLEM Relatively fev tests have been devised to measure a b i l i t y i n soccer. Most of the existing tests are i n battery form, with t h e i r concomitant disadvantages of requiring too much space, time, equipment and supervision. While some battery-type tests have provided a reasonable basis f or predicting or measuring soccer a b i l i t y , they have enjoyed l i t t l e acceptance among teachers, coaches and physical educators. Voltmer and Esslinger ( l ) state, "A vast majority of physical educators today do not share t h i s enthusiasm fo r testing and are prone to r e s i s t attempts to set up such programmes." Their chief c r i t i c i s m s are that many of the available tests are neither r e l i a b l e nor v a l i d and that a testing programme to meet t h e i r needs would involve too much time and, i n many cases, too much expense. However, teachers and physical educators do prefer to objectively evaluate th e i r work when possible, and i t i s with t h i s i n mind that t h i s test has been developed. 5 REFERENCES 1. Voltmer, E. F. and Esslinger, A.A., The Organization and  Administration of Physical Education. 3rd Ed. (Appleton-Century-Crofts, Inc., New York, 1958, p.505. CHAPTER III REVIEW OF LITERATURE Evelyn F. Schaufele ( l ) studied the value of objective tests f o r g i r l s i n determining soccer a b i l i t y . Her tests were given to a group of eighty-four g i r l s i n the ninth and tenth grades i n the schools of Fairview V i l l a g e , Ohio, i n the f a l l of 1938. C r i t e r i a were secured by having each g i r l rated subjectively by three instructors and two senior students, the observations being carried out during two class periods. A conclusion was that the best single test f o r measuring soccer a b i l i t y of high school g i r l s was that of wall-volleying, which had a correlation with the c r i t e r i a of .57. Schaufele also correlated her wall-volleying test with the sum of several other tests which had been T-scored. Her figures shoved that the test ranked t h i r d i n correlation with subjective c r i t e r i a and f i r s t with the combined tests c r i t e r i a having a relationship of .77 by r e c t i l i n e a r correlation. Schaufele's wall-volley consisted of rebounding a b a l l against a wall 15' vide by 10' high. She used an i n i t i a l restraining l i n e of 15 feet, but once the test had begun, the restraining l i n e did not enter into the test; however, she did set l i m i t s on the forecourt area which was 30 feet square. Each subject had two t r i a l s of one minute duration and only the best score was counted. Marjorie L. Heath and Elizabeth J . Rogers (2) made a study on the use of knowledge and s k i l l tests i n soccer, using as subjects f o r the experiment unskilled children i n the f i f t h and sixth grade. 7 The s k i l l tests were: 1* Soccer dribble t e s t . 2. Throw-in for accuracy. 3. Kicking a dead b a l l f o r distance. 4. Scoring goals, kicking a dead b a l l from the penalty spot. 5. Knowledge test s . Results were: 1. In grade V the correlation between the composite T-score of the soccer s k i l l t est and the judgement rating was .602. 2. In grade VI the relationship between the composite T-score and judged playing a b i l i t y was .624. Bontz (3) developed a dribble and shoot t e s t f o r f i f t h and sixth grade children. The test was given to 142 g i r l s . In a review of the l i t e r a t u r e and from questionnaire returns she established that kicking constituted the basic s k i l l of soccer and the s k i l l most often tested and taught. Her test proved rather unwieldy as she required four successful runs f o r : dribbling, passing, and scoring. The c h i l d used f i r s t the l e f t foot and then the right foot before the test was completed. (See diagram No. l ) . The test suffered because of the necessity of successful completion, and involved too much time. The v a l i d i t y of the test was reported as .58 by correlating player's times with player's ratings. The r e l i a b i l i t y coefficients were calculated by using the odd and even performances of the player's l e f t and right fe e t . These were given as .85 and .91 respectively. 8 Rebound Wall GOAL 4 >> Dribble, pass off wall, shoot through goal Diagram I Crawford (4) revised the Schaufele wall-volley t e s t f o r use with women majoring i n Physical Education at the University of Oregon. Crawford did not bother with a restraining l i n e , but supplied a retriever service to retrieve the b a l l i n the event of miskicks, poor kicks and misses. Crawford found a v a l i d i t y c o e f f i c i e n t of only .252 by correlating the rating of three judges with the i n i t i a l test results. Crawford had more success with a rebound and trap test which she devised. The b a l l had to be rebounded from a wall and trapped behind an 8 foot restraining l i n e . She reported a v a l i d i t y c o e f f i c i e n t of .450 between the best score of three t r i a l s and the judges* ratings, and a v a l i d i t y c o e f f i c i e n t of .537 between the best score of the three t r i a l s and the t o t a l test c r i t e r i a . The test r e l i a b i l i t y was .704 using the second and t h i r d t r i a l scores. In 1950 Konstantinov (5) experimented at Springfield College with the purpose of developing and evaluating a battery of soccer s k i l l s as an index of a b i l i t y i n the game of soccer. His tests, developed by means of expert opinion and composite score c r i t e r i a , were to be used f o r both c l a s s i f i c a t i o n and diagnosis. Data were 9 obtained from the results of testing of seventy-four vars i t y , freshmen varsit y , and soccer class members. He used factor analysis to reveal that as f a r as the tests used i n his study were concerned, there were three fundamental factors related to soccer - s k i l l , power, and speed. Vanderhoof (6) devised a battery test consisting of ten items: dribbling, trapping, throwing, tackling, place-kicking, volleying, corner-kicking and goalkeeping. She did not give any figures of student performance but suggested that the tests be considered as measures, or be evaluative, of players performance i n each of these areas. Winterbottom (7) t r i e d to test three s k i l l s : place-kicking using a moving b a l l , accuracy i n kicking using a moving b a l l , and controlled heading a b i l i t y . He tested s i x t y of the top professional soccer players i n England, but the average results given by him for f i v e kicks with the l e f t foot i n each category (2 out of 5) would indicate that either the test i t s e l f was too d i f f i c u l t , or the players were not motivated to perform to the best of t h e i r a b i l i t y . This was also true of h i s two heading tests which produced low average results of one i n f i v e . MacDonald (8) experimented with wall-volley testing by reducing the restraining l i n e distance from an i n i t i a l t r i a l distance of 30 feet to one of 9 feet. MacDonald*s target area was 30 feet wide and ll£ feet high. The players started the test with a stationary b a l l from behind the restraining l i n e and two spare b a l l s were placed 9 feet behind the restraining l i n e i n the centre of the testing area. 10 Use of the spare b a l l s allowed the subject to make a recovery i n the event of his losing the o r i g i n a l b a l l . MacDonald allowed the subjects to control the rebounding b a l l behind the restraining l i n e i n any manner possible including use of the hands. In the event of the subject retri e v i n g a poorly h i t b a l l or selecting a spare b a l l , he was permitted to use his hands to return the b a l l to the starting position behind the restraining l i n e and continue the t e s t . Each subject was allowed four t r i a l s of 30 seconds each. The score was the highest score of any three t r i a l s . MacDonald tested three groups of v a r s i t y players engaged i n soccer. He used the subjective ratings of the three coaches of the three groups against t h e i r performance scores i n computing the v a l i d i t y of the t e s t . He obtained the following coefficients of correlation. Correlation with  Number Group Subjective Rating 17 Varsity Team .94 18 Junior Varsity .63 Players 18 Freshmen Players .76 53 Combined Groups .85 Mitchell (9) used a revision of the MacDonald wall-volley t e s t to determine the s u i t a b i l i t y of wall-volley testing as a technique for evaluating the soccer a b i l i t y of grade f i v e and s i x elementary school boys i n West Vancouver, B r i t i s h Columbia. A restraining l i n e of 6 feet and a target area of eight feet long by four feet high was used. Three t r i a l s of 20 seconds, and use of the spare b a l l s technique 11 and a retriever service were employed. The subjects started the test by kicking a stationary ball from behind the restraining line. No use of the hands was permitted in recovering a badly played ball . Mitchell correlated a coaches* rating with test and re-test performance scores. The test and re-test were administered to six groups on the same day. Mitchell used combinations of t r i a l scores and correlated these with the coaches' ratings. The following validity coefficients of correlation using two methods were obtained. Method Groups Mean 1 2 3 4 5 6 Rank Difference .864 .859 .821 .846 .825 .841 .84 Product Moment .768 . 808 . 699 . 813 .748 .717 .76 The wall-volley testing technique has also been used to measure ability in other sports such as tennis, badminton, volleyball and handball. Miller (10) made use of the wall-volley technique to devise a badminton test as a measure of badminton ability. A restraining line at 10 feet and a rebounding area extending above a line drawn on the wall at a height of l\ feet. Three 30-second volleys or rallies were allowed and the score consisted of the sum of the three t r i a l s . The test r e l i a b i l i t y was determined by the test-retest method and found to be .94, while the test validity of .83 was determined by correlating the test scores with the results of a round-robin tournament. 12 Dyer ( l l ) used a similar technique i n setting up a test to measure tennis a b i l i t y and also as a means of c l a s s i f y i n g subjects. A restraining l i n e of 5 feet from the base of the backboard which was 10 feet high by 15 feet wide, with a l i n e drawn at a height of 3 feet to represent the net. A box of extra b a l l s was provided at the side of the restraining l i n e . The t o t a l score of three 30-second t r i a l s was taken. The t e s t r e l i a b i l i t y was found by the test-retest method to be .86, while the v a l i d i t y of the test was determined by correlating the test scores with subjective judgements of three experts, and also by correlating the best scores with standings obtained by the subjects i n several round-robin tournaments. The f i r s t of these methods revealed a v a l i d i t y c o e f f i c i e n t of correlation of .85, while the second method produced coefficients ranging from .85 to .92. Brady (12) made use of the wall-volley as a means of cl a s s i f y i n g and grading college men i n v o l l e y b a l l . Brady experimented with several test items but found the wall-volley test to be most v a l i d . No restraining l i n e was used, but a smooth rebounding area was necessary. A l i n e 5 feet long and l l i feet high was drawn on the wall, and v e r t i c a l l i n e s extended towards the c e i l i n g at the ends of the horizontal l i n e . The subject had to volley the b a l l f o r one minute. The test was begun by throwing the b a l l against the rebounding area. Only l e g a l volleys counted. The test-retest method revealed a r e l i a b i l i t y of .925 (for 282 subjects), while the test v a l i d i t y of .86 was determined by correlating the test scores against the subjective judgements of 4 judges. 13 Russell and Lange (13) also made use of a wall-volley test to measure volleyball ability i n junior high school g i r l s . Their test was really a modification of an earlier test developed by French and Cooper (14). They used a restraining line of 3 feet, and a rebounding area 10 feet wide and 7i feet high. The score was the total number of legal hits i n three 30-second t r i a l s . The test r e l i a b i l i t y , using the test-retest method, was found to be .87, while the test validity of .80 was obtained by correlating the test scores of the subjects with the subjective ratings of seven judges. French and Cooper used the same technique, except that they allowed ten 15-second t r i a l s . The score taken was the sum of the five best t r i a l s . Cornish (15) devised a series of tests to measure handball ability including the 30-second wall-volley. The ball was served from the service zone. The total number of rebounds across the service line was counted. In the event of a ball getting out of control, a judge handed the contestant another ball . Combining the 30-second volley with the Service Placement test provided the best coefficient (.667) when correlated with the criterion (the subject's total points score in relation to his opponents after 23 games). 14 REFERENCES 1. Schaufele, Evelyn, F., "The Establishment of Objective Tests for Girls of the Ninth and Tenth Grades to Determine Soccer Ability". (Unpublished Master's Thesis Abstract, Springfield College, 1951). 2. Heath, Marjorie,L. and Rogers, Elizabeth, G., "A Study in the Use of Knowledge and S k i l l Tests in Soccer", Research  Quarterly, vol. 33, no. 3, December, 1932. ^ 3. Bontz, Jean, "An Experiment in the Construction of a Test for Measuring Ability in Some of the Fundamental Skills Used by Fifth and Sixth Grade Children in Soccer". (Unpublished Master's Thesis, State University of Iowa, 1942). 4. Crawford, Elinor, A., "The Development of S k i l l Test Batteries for Evaluating the Ability of Women Physical Education Major Students in Soccer and Speedball". (Unpublished Doctoral Thesis, University of Oregon, 1958). 5. Konstantinov, J., (Unpublished Master's Thesis, Springfield College, 1950). 6. Vanderhoof, Mildred, "Soccer S k i l l Tests", Journal of Health and Physical Education, vol. 42, no. 3, October, 1932. 7. Winterbottom, Walter, Soccer Coaching, revised ed., Naldrett Press Ltd., London, I960, pp. 243-244. 8. MacDonald, Lloyd, G., "The Construction of a Kicking S k i l l Test as an Index of General Soccer Ability". (Unpublished Master's Thesis Abstract, Springfield College, 1951). 9. Mitchell, Reid, "A Wall-Volley Test for Measuring Soccer Ability in Fifth and Sixth Grade Boys". (Unpublished Master's Thesis, University of Oregon, 1963). 10. MacDonald, op.cit. H i Dyer, Joanna, T., "Revision of Backboard Test of Tennis Ability", Research Quarterly, vol. 9, no. 1, March, 1938. 12. Brady, George, F., "Preliminary Investigations of Volleyball Playing Ability", Research Quarterly, vol. 16, no. 1, March, 1945. 13. Russell, Naomi, and Elizabeth Lange, "Achievement Tests in Volleyball for Junior High School Girls", Research  Quarterly, vol. 2, no. 4, December, 1940. 15 14. French, Esther, L., and Cooper, Bernice, "Achievement Tests in Volleyball for High School Girls", Research Quarterly, vol. 8, no. 2, May, 1937. 15. Cornish, Clifford, "A Study of Measurement of Ability in Handball", Research Quarterly, vol. 20, no. 2, May, 1949. CHAPTER IV METHODS AND PROCEDURES An appropriate test of soccer s k i l l should approximate as nearly as possible the elements present in the actual soccer game: shooting, passing, dribbling, tackling, trapping and heading. A game-like situation requires that the aforementioned elements be performed whilst the ball i s i n motion and an "under pressure" situation in existence. Many tests f a i l to meet either of these criteria. Finney ( l ) , Meisl (2), and Czaknady (3), who played representative soccer for their respective countries i n international competition, each l i s t the aforementioned s k i l l s as prerequisites of the accomplished soccer player. A poll of the opinions of the top four clubs in the Vancouver, B.C. Pacific Coast Soccer League (Canadians, Columbus, Firefighters and Victoria United) revealed that they rated the six foregoing sk i l l s above a l l others in determining soccer ability. Ve may visualize each of these elements entering into a wall-volley test situation: shooting to make the ball strike the target area sooner; passing to make sure that the ball strikes the target area; trapping by controlling an awkwardly bouncing ball; tackling by intercepting and controlling a di f f i c u l t return; dribbling by returning a rebounding ball quickly to the restraining line, and, of course, heading a high return. In a review of previous tests which attempted to measure soccer s k i l l , the wall-volley-type test came closest to approximating 17 "under pressure" conditions of the game situation* The changes introduced i n th i s study were: (a) The size of the target area - For l o g i c a l reasons, the target area decided upon was the same as the regulation goal measurement (24 feet by 8 f e e t ) . This size was chosen as one that would be familiar to a l l players, and an area towards which a l l players would be accustomed to playing a b a l l . (b) The distance of the restraining l i n e - The restraining l i n e distance of 15 feet was decided upon after experimentation with restraining l i n e s of 30 feet and 24 feet. (c) Starting the test by kicking a moving b a l l - Previous wall-volley tests had used a dead b a l l s t a r t behind the restraining l i n e . Other battery tests had used a b a l l thrown or r o l l e d by the researcher. I t was f e l t that the b a l l should not be dead, nor should i t be subject to variations i n velocity, height, bounce or angle of a r r i v a l , so i t was decided that the player should hold the b a l l i n his hands behind the restraining l i n e . On the command "go" he should put i t into play as quickly as possible, thus, leaving the control of the b a l l e n t i r e l y to the indiv i d u a l . (d) The use of spare b a l l s - Subjects were not to be penalized f o r using spare b a l l s . (e) The method of scoring - The aggregate score to count rather than the best t r i a l or best two t r i a l s . A rank ordering of player technique was adopted and rated 18 by the investigator. Weiss and Scott (4) state "Judges ratings can be a satisfactory criterion i f the judges are Competent and well trained, and i f they have an adequate chance to observe before rating." In an attempt to overcome any element of subjectivity which may have been occasioned by the system of rank ordering, the players were rated after observing them play twelve soccer games with their respective teams. In assessing the rating of players whom the experimenter subjectively tie-ranked, a "checklist" technique was used. Players who could shoot hard and accurately with both feet were rated higher than those who could only shoot well with one foot. Players who could pass accurately were preferred to those who gave careless or incompleted passes. Players who made interceptions or successful tackles were rated above those who were poor at intercepting or tackling. Players who completed successful dribbles when forced to do so were preferred to those who were not successful. Players who revealed good trapping control were rated above those who did not, and players who were successful in heading a ball when challenged by an opponent were rated superior to those who were not. The test was constructed so that i t would demand the effective combined use of the six basic sk i l l s of shooting, passing, trapping, tackling, heading and dribbling, a l l performed "under pressure". 19 DIAGRAM 2 Target Area 8« 24» Student counting X " f a u l t s " R e s t r a i n i n g Line I 15' A i Student counting -X "hits" X Experimenter-Subject with ball InT KandsT to begin test Box of \ i / ^spare balls 151 Test Construction: (a) The target area was set at regulation goal size (24* x 8* high). (b) The restraining line was fixed at 15' and parallel to the base of the target area. (c) The box of spare balls was kept at a distance of 15* behind the centre of the restraining line. (d) The decision to use a spare ball was l e f t to the player (subject), (e) The subject to start the test stood behind the centre of the restraining line facing the target area, holding a ball in both hands at waist height. (f) The aggregate score of three 30-second trial s counted as the subject's score. 20 (g) The experimenter gave the commands "go" and "stop" and took the time on a calibrated stop watch. (h) A student counted the total number of hits to cross the restraining line during each t r i a l . ("Hits": any ball played correctly by the subject from behind the restraining line, striking the target area and rebounding across the restraining line). (i) A student counted the number of "faults" made by the subject during each t r i a l , and these were deducted from the total number of rebounds noted by the f i r s t student. ("Fault": any infringement of a "hit" made by a subject within the t r i a l limit). (j) A third student provided a retriever service by returning a ball abandoned by the subject to the spare ball box. (k) Only regulation size rubber soccer balls, inflated to 12 lbs. pressure, were used. The wall-volley test was developed by the experimenter to serve as a means of measuring soccer s k i l l , and also as a method of grading or classifying large groups of players. The test illustrated that five distinct categories of soccer ability may be revealed: superior, good, average, below average, and poor. While in this study the test was administered to university students, i t is believed that i t s administration will separate subjects of a l l ages into these five categories. Administration of the Test: 1. The test consists of three 30-second trial s of rebounding a 21 soccer ball from behind a restraining line at a distance of 0 15* against a target area of 24' long and 8' high. The forecourt surface is tarmacadam and dry. 2* To start the test the subject stands behind the restraining line facing the target area and holds the ball i n both hands at waist height* 3. A box of spare balls i s provided 15 feet behind the centre of the restraining line. 4. On the command "go" the subject drops the ball from his hands and commences to rebound the ball against the target area as often as possible within the 30-second t r i a l period* 5. The ball may be directed by the foot, leg, knee, or other part of the body except the hands or arms, to the area marked on the target area. 6* A ball that does not rebound across the restraining line can be retrieved by the subject by dribbling i t across the restraining line, or by rebounding i t from the target area across the restraining line. In either case the incompleted rebound would not count* 7. The use of the hands at any time to steady or to retrieve the ball i s not allowed* 8. On the command "stop" at the end of 30-seconds, the subject's score i s the total number of correctly completed rebounds. 9. A retrial i s to be given the subject in the event that the retriever (student) interferes i n any way with the subject's performance* 22 10. The three t r i a l s are to be carried out v i t h a minimum of delay between t r i a l s . Instructions to Subjects: 1. On the command "go", s t a r t the test immediately. Drop the b a l l ; i t need not bounce before you play i t against the target area. Continue to play the b a l l to the target area u n t i l the command '•stop" at the end of 30-seconds. 2. Tou may use any s k i l l or combination of s k i l l s . l o u must play a l l b a l l s from behind t h i s restraining l i n e (indicate the l i n e c l e a r l y ) . 3. l o u may cross the l i n e to retrieve the b a l l , but any " h i t s " made i n such a position do not count. You may use any number of b a l l s . I f f o r any reason you lose close contact with the b a l l i n play, do not t r y to retrieve i t . Take another b a l l from t h i s box (indicate clearly) and put i t i n play as you did at the s t a r t . 4. Each b a l l s t r i k i n g the wall i n the marked area and returning over the restraining l i n e before the word "stop" counts as a h i t and scores one point. 5. You w i l l each be given three t r i a l s today. The f i n a l score on the test i s the sum of the scores on the three t r i a l s . The following points are to be demonstrated: 1. One b a l l i n hands. 2. Start test by dropping b a l l , then play i t . 3. Demonstrate a few times, showing various s k i l l s : side-foot, instep, knee, thigh and head v o l l e y . 23 4. Cross restraining l i n e to retrieve a b a l l , make a low h i t to keep i t i n play, and retreat f or next shot* 5. Make a wild shot to show how taking another b a l l saves time. Put this new b a l l i n play as at the st a r t . Read the following paragraph to make certain that each person understands the tes t procedure and his duties. No. 1 takes the t e s t . At the signal "ready" he stands i n the centre behind the restraining l i n e facing the target area with a b a l l i n his hands, prepared to start the tes t at the word "go". No. 2 counts the number of b a l l s which strike the wall i n the marked area and recross the restraining l i n e before the word "stop", and enters them on the score card opposite the approximate t r i a l number. I f any infringements are reported by No. 3, these are deducted before the score for the t r i a l i s recorded. A ball'recrossing the restraining l i n e coincident with the word "stop" counts. No. 3 watches the player i n r e l a t i o n to the restraining l i n e . He reports to the scorer at the end of the t r i a l the number of h i t s , i f any, made while the player was standing closer to the wall than the restraining l i n e . No. 4 col l e c t s the b a l l s before the start of a t r i a l and puts them i n the box. During the t r i a l he retrieves and returns to the box any b a l l s going out of play. Method of Scoring Dr. McCloy (5) suggested i n the revision of the backboard tes t of tennis a b i l i t y that the aggregate t o t a l should be each player's score rather than simply taking the best of three t r i a l s e 24 He suggested that the time lost in putting another ball in play ... might cause sufficient reduction in score without any additional penalties, such as subtracting the number of extra balls used from the number of hits scored. This method was adopted in developing the soccer test described in this study. Gathering the Data Five groups of students playing soccer at the University at the time of this study were selected as the experimental groups. The five groups were the three university representative soccer teams: The Thunderbirds, the Chiefs and the Braves; a Physical Education Major soccer class, and a Required Programme soccer class. The experimenter, through a programme of visitation during class periods, observed these groups playing soccer on no fewer than twelve occasions. Following this period of observation, a rank order was determined for the members of each group. Players of each group were ranked in order of ability, from one to fifteen by the experimenter. Goalkeepers were excluded from rank orders. Following the administration of the test the players were rank-ordered again from one to fifteen on the basis of their scores. The groups were tested within a one-week period between 1:00 and 2:00 P.M. on separate dry days, thus standardizing ground and atmospheric conditions. A l l players wore running shoes. The test was conducted on a blacktop area against a smooth cement wall 8 feet high and 24 feet wide. This area (same dimensions as a goal area) was clearly marked. A restraining line was drawn 15 feet from, and parallel to, the base 25 of th© v a i l , and was c l e a r l y v i s i b l e . Treatment of Data The test scores of each group were rank-ordered from one to f i f t e e n , or from highest to lowest. Using the rank difference method of correlation, the test scores of each group were compared with the experimenter's pre-test rating of the players i n each group. The degree of relationship between these two rank orders was determined by the c o e f f i c i e n t of correlation obtained by using the Spearman Brown Rank Order formula: r = 1 - 6 x d 2 N(N-l) Thus, f i v e rank difference coefficients of correlation were obtained; one for each group. A t o t a l group v a l i d i t y c o e f f i c i e n t of correlation vas obtained by using the formula f o r q u i n t i s e r i a l correlation as outlined by Jaspen (6): quint r = ZaYa(Zb-Za) Yb-t-(Zc-Zb) Yc+(Zd-Zc) Yd-ZdYe 6y Za 2+ (Zb-Za) 2+ (Zc-Zb) 2+ (Zd-Zc) 2+ Zd 2 a b c d e The r e l i a b i l i t y of the test was found by comparing the scores obtained on each of the three t r i a l s . Thus, three r e l i a b i l i t y c orrelation coefficients were obtained, i . e . by comparing the scores of the f i r s t t r i a l with the scores obtained i n the t h i r d t r i a l , by comparing the scores obtained on the second t r i a l with those obtained on the t h i r d t r i a l , and l a s t l y , by comparing the scores 26 obtained on the f i r s t t r i a l with those obtained on the second t r i a l . To obtain these coefficients of correlation, the following two formulas were used: (a) The Pearson Product Moment Formula, and (b) The Spearman Brown Prophecy Formula. Garrett (7) states "Increasing the length of a test or averaging the scores obtained from several applications of the test or from parallel forms will increase r e l i a b i l i t y . Fortunately a good estimate of the effect of lengthening or repeating a test can be obtained by use of the Spearman Brown Prophecy formula." The standard deviation, the mean, the median and the aggregate scores were also calculated for each group and also for the total group of 75 subjects. 27 A L-REFERENCES 1. Finney, Thomas, Instructions to Young Footballers. Museum Press Ltd., London, 1955, pp. 16-34. 2. Meisl, W., Soccer Revolution. Phoenix Sports Club, London, 1955, pp. 20-34. 3. Czaknady, Jeno, Learn to Play the Hungarian Way. Hungarian Sport Publishing House, Budapest, 1954, pp. 17-71. 4. Weiss, Raymond, A., and Scott, Gladys, M., "Construction of Tests". Research Methods. 2nd ed., American Association for Health, Physical Education and Recreation, Washington, D.C., 1959, pp. 239-240. 5. Dyer, Joanna, T., "Revision of the Backboard Test of Tennis Ability", Research Quarterly, vol. 9, no. 1, March 1938, p. 27. 6. Jaspen, Nathan, "Serial Correlation", Psychometrika, March, 1946, pp. 23-30. 7. Garrett, Henry, E., Statistics in Psychology and Education. 5th ed., Longmans, Green, and Co., New York, 1958, p. 343. CHAPTER V PRESENTATION OP DATA The five groups - the Thunderbirds, the Chiefs, the Braves, the Physical Education Major Soccer Class, and a Required Programme Soccer Class were a l l given the wall-volley test as described i n Chapter 3. A total of 75 subjects were tested. There were 15 subjects in each group and each group was rank-ordered by the experimenter from 1-15. The scores of each subject and of each group are indicated in Tables 1 to 5 of the appendix. The range of scores of each group was: Thunderbirds Chiefs Braves P.E. Majors Required Programme 58 - 40 50 - 27 47 - 25 34 - 21 36 - 16 (18) (23) (22) (13) (21) TABLE 1 The mean, median and standard deviation scores of each group were: Group Mean Score Median Score Standard Deviation Thunderbirds 48.00 47.00 4.99 Chiefs 36.93 38.00 6.26 Braves 32.60 31.00 6.34 P.E. Majors 25.66 24.50 4.59 Req'd. Programme 24.20 23.50 5.76 Coefficients of correlation for the test validity were obtained for the five groups by using the rank order difference 29 method of computation. The following validity coefficient of correlations were obtained for each group. Thunderbirds Chiefs Braves P.E. Majors Required Programme .577 .840 .812 .944 .975 A validity coefficient of correlation for the total group tested was obtained by using the quintiserial technique (Jaspen) (l) of obtaining a coefficient of correlation. The quintiserial method of computing a correlation coefficient was obtained by comparing the five group categories with the five group test scores. Using this technique, the total group test validity was found to be .856. The r e l i a b i l i t y coefficients of correlation were found by' comparing the scores obtained by the subjects in the three t r i a l s . Thus, three r e l i a b i l i t y coefficients of correlation were found: one between the f i r s t and the third t r i a l ; one between the second and third tria l s and one between the f i r s t and second t r i a l s . These r e l i a b i l i t y correlation coefficients were obtained by means of the Pearson Product Moment Formula and the Spearman Brown Prophecy Formula. The following r e l i a b i l i t y correlation coefficients were obtained: 1st & 3rd Trials 2nd & 3rd Trials 1st & 2nd Trials .900 924 921 30 The r e l i a b i l i t y of the test had already been established i n previous studies of wall-volley t e s t s . Garrett (2) states " a highly v a l i d test cannot be unreliable since i t s correlation with a c r i t e r i o n i s limited by i t s own index of r e l i a b i l i t y . " Standard deviations for the f i v e test groups are included i n Table 1. The standard deviation f o r the whole group was found to be 10.33. A sixth group of subjects composed of 16 members of another Physical Education major soccer class were given the test on successive weeks to determine: (a) the effects of practice (b) when peak performance was reached. This sixth group was not rank-ordered, but merely given the test with the same te s t instructions, and under the same conditions. Over a five-week period t h e i r aggregate scores were: Sub.i ect 1st T r i a l 2nd T r i a l 3rd T r i a l 4th T r i a l 5th T r i a l A 26 28 37 43 30 B 24 18 27 40 30 C 22 33 32 26 39 D 34 36 36 45 32 E 28 29 24 34 38 F 31 41 49 36 — G 22 25 38 34 51 H 36 40 48 44 36 I 36 37 42 48 34 J 25 29 26 28 27 E 26 35 28 32 40 L 34 39 32 42 32 M 17 25 28 34 — N 10 8 14 21 25 0 28 28 38 33 33 P 24 25 — . — — 31 Table 2 shows the mean, the range, and the standard deviation scores i for the sixth group over five tria l s * TABLE 2 Trial Number Mean Scores Range of Scores Standard Deviation 1 27.06 10 - 36 6.85 2 29.75 8-41 7.64 3 33.27 14 - 49 8.99 4 36.00 21-48 7.39 5 31.93 25 - 41 6.46 TABLE 3 Smoothing Frequency Distribution Curve it of Interval Frequency Correction; 13.5 0 .66 16.5 2 1.33 19.5 2 4.66 22.5 10 6.66 25.5 8 9.33 28.5 10 8.66 31.5 8 8.00 34.5 6 6.00 37.5 4 5.66 40.5 7 5.00 43.5 4 5.66 46.5 6 4.66 49.5 4 3.66 52.5 1 2.33 55.5 2 1.33 50.5 1 1.00 61.5 0 .33 "To find an adjusted or "smoothed" f, we add the f on the given interval and the f's on the two adjacent intervals and divide the sum by 3." (3) For example, the smoothed f ffir mid-point 22.5 is 10 + 2 + 8 - 20 = 6.66 3 3 32 TABLE 3 Step Cumulative Percentile Interval Frequency Frequency Score 57 - 59.9 1 75 P,o*= 60 54 - 56.9 2 74 P*; = 51.75 51 - 53.9 1 72 Pic = 48.17 48 - 50.9 4 71 Pgi = 46.37 45 - 47.9 6 67 Tgo = 43.8 42 - 44.9 4 61 P« = 41.68 39 - 41.9 7 57 P 7 o = 40.25 36 - 38.9 4 50 Tts = 38.06 33 - 35.9 6 46 Pfc> = 35.14 30 - 32.9 8 40 P« = 33.62 27 - 29.9 10 32 T/0 = 32.75 24 - 26.9 8 22 "Bus = 30.66 21 - 23.9 10 14 ?4» = 29.4 18 - 20.9 2 4 P» =28.27 15 - 17.9 2 2 P*>= 27.15 Pa = 25.78 75 P^ = 24.35 Pi* = 23.90 P<"= 22.05 P/ = 20.62 Po = 14.90 Presentation of Data The curve of the frequency d i s t r i b u t i o n of the t o t a l group was smoothed by using the technique as explained by Guilford (4). As the di s t r i b u t i o n did not reproduce a normal curve, i t was decided to show percentile scores rather than transform scores into standard scores. 33 REFERENCES 1. Jaspen, Nathan, "Serial Correlation", Psychometrika. March, 1946, pp. 23-30. 2. Garrett, Henry, E., Statistics in Psychology and Education. 5th ed., Longmans, Green and Co., New York, 1958, pp. 471. 3. Ibid.. P. 13. 4. Guilford, J . P., Fundamental Statistics in Psychology and Education. 2nd ed., McGraw H i l l Book Co., New York, 1950, pp. 51-54. CHAPTER VI ANALYSIS OP DATA The test results show that the experimenter correctly-categorized the five groups as being representative of five different levels of soccer s k i l l displayed by students at the University of British Columbia. Substitution of the five group names by the five classes of superior, good, average, below average, and poor would seem to be quite justified. On the basis of the numbers tested, and of the results obtained in this study, a table of norms is suggested for use in classification. Superior 42 and over Good 37 - 41 Average 31 - 35 Below Average 25 - 30 Poor 24 and below The five group validity coefficients of correlation show that greatest difficulty in rank-ordering was experienced with subjects who participated i n the higher classes of soccer (superior and good). Mathews (l) states that validity coefficients from .80 -.85 may be interpreted as very good, and above .85 as excellent; those falling within the range .70 - .79 may be considered acceptable, especially where a subjective judgement is involved. Garrett (2) states that "the validity of a test depends upon the f i d e l i t y with which i t measures what i t purports to measure." 35 The validity of this test has been determined experimentally by finding the correlation between the test and the external criterion (the experimenter's rating). Group validity coefficients of correlation which f a l l in the range between .737 and .975 are quite acceptable. The only validity correlation coefficient which f a l l s below acceptable standards is that obtained for the Superior or Thunderbird Groups, .577. This is understandable considering the difficulties that present themselves in ranking subjects who play in the superior category, and whose test scores revealed that two-thirds of the group f e l l between 5 scores above and 5 scores below the mean of 48. It i s obvious that the Superior or Thunderbird Group is quite distinctly differentiated in performance from a l l other groups. The groups representing the "below average" and "poor" categories are much less different from each other than they are different from the average, good and superior groups. It was assumed that the Physical Education majors group would be superior to the Required Programme group because they were older, and as Physical Education students might be expected to reveal greater interest, coordination and sports s k i l l ability. This difference, i t was f e l t , might be more pronounced i f the test was administered to greater numbers of students in those respective groups. The validity coefficient of .856, obtained by using the quintiserial correlation method, is considered excellent according to the criteria described by Mathews (3). This validity coefficient 36 is higher than that of .57, obtained by Schaufele (4), and the .76 obtained by Mitchell (5) and very similar to MacDonald's (6) .85. It should also be noted that each of these researchers obtained their test validity coefficient figures by taking a mean of the group validity coefficients. The difference of 4.33 between the mean scores of the Chiefs (good) and the Braves (average) groups, while significant, might reasonably have been expected to be greater. The season (1962-63) of Varsity soccer has produced an unbeaten Thunderbird team, an extremely weak Chiefs team, and a rather strong Freshmen, or Braves representative team. One would normally have anticipated more equality in the intervals between the means of the f i r s t three groups. These peculiarities of the 1962-63 season saw the Thunderbird (superior) group win league and cup tournaments, while the Chiefs (good) group finished at the bottom of their division. The Braves (average) group won promotion to a higher division which probably contributed quite considerably to the inequality of the intervals between the mean scores of the groups. However, this unevenness in playing s k i l l might occur by chance i n any given year in Varsity representative soccer teams. In view of the difficulties described, i t is fortunate that the groups did reveal distinct differences in mean performance score, in aggregate performance score, and in group validity correlations. The validity correlations of the five groups were produced by means of the rank difference correlation method and were a l l acceptable except that of .577 found for the Thunderbird group. 37 These correlations express the degree of relationship between the experimenter's rank ordering of the subjects i n terms of soccer a b i l i t y , and the rank order of the subjects according to t h e i r t o t a l performance score on the t e s t . The experimenter was a former professional soccer player with the Glasgow Rangers for seven yearsj a Canadian A l l - S t a r and current player coach of the University of B r i t i s h Columbia Thunderbird team. This suggests that the background of experience and part i c i p a t i o n q u a l i f i e s him as an expert capable of assessing the a b i l i t y of players after having had the opportunity to observe them at play on no less than 12 occasions. The Required Programme group rank order correlation of .975 was almost a perfect correlation because the experimenter was able to detect quite d i s t i n c t differences i n the individual a b i l i t i e s of these subjects. This was possible because for many of these subjects t h i s was t h e i r f i r s t experience i n soccer, and the various differences i n s k i l l and a b i l i t y were more c l e a r l y defined i n this group than i n any other. These variations i n s k i l l and a b i l i t y , while quite c l e a r l y evident to the trained observer i n the Physical Education Major and Required Programme groups, became progressively less differentiated i n the representative teams. Perhaps a possible source of error i n rank-ordering the Thunderbird group was that the experimenter was thoroughly familiar with the player's previous soccer experience, and was also a fellow player with them. Thus, most of the observations of t h i s group were done by the experimenter as a participant, while those of the other groups were done while the experimenter was i n the role of a spectator. 38 The fact that the Thunderbirds (superior) group was a homogeneous group of high ability, made rank-ordering d i f f i c u l t , while the Physical Education Majors (below average) group was a homogeneous group of low ability, yet easily rank-ordered. This would suggest that group homogeneity alone does not preclude rank-ordering, but rather homogeneity plus high s k i l l performance makes rank-ordering an unrealistic technique. The Chiefs (good), the Braves (average) and the Required Programme (poor) groups were heterogeneous groups and readily lent themselves to a rank-ordering technique. The superior group reveals a small range (18); the score distributions form a normal curve about the mean and the mean -1 standard deviation encompasses two-thirds of the scores. Another significant feature of the rank ordering by the experimenter is that in a l l cases the f i r s t ranked occupied the same position in the performance rankings. Thus, choosing the best player did not pose diffic u l t i e s . The greatest area of difficulty in ranking, in groups other than the Thunderbirds group, presented it s e l f in the eleventh to fourteenth positions. While i t was possible to predict the last ranked, or lowest performer, almost as accurately as the f i r s t ranked, i t was rather d i f f i c u l t to distinguish between the subjects beyond the tenth rank order. It was, therefore, not surprising to find many tie-rankings in the performance scores between the eleventh and fifteenth rank-ordered positions. If the mean of 34 for the whole group is accepted as the "pass" performance score, then only 33 subjects obtained a "pass" while 42 subjects failed. However, this average has been boasted 39 by the high performance scores of the Thunderbird group. The medium score for the total group lies between 30 and 29, and this would appear to be a more suitable "pass" mark in this instance. The test r e l i a b i l i t y was determined by correlating the scores obtained on the three t r i a l s . The correlations of .742 between the f i r s t and second t r i a l s , .801 between the second and third t r i a l s , and .796 between the f i r s t and third t r i a l s , are quite acceptable in this form of test, especially when one considers that only 75 subjects were used in the test and that the t r i a l s were performed one after the other without any noticeable rest period in between. If the test were administered on three separate occasions, i f three parallel forms of the test were administered, then "the r e l i a b i l i t y of the averaged scores will be the same as the r e l i a b i l i t y obtained by tripling the length of the test." (7) Thus, these re l i a b i l i t i e s could be increased by tripling the length of the test to .900, .924, and .921 by using the formula, //Tv = n r l l 'f^ 1 + (n-l)r H where Jfi^ = the correlation between n forms of the test and n alternate forms, r„ = the r e l i a b i l i t y of Trial 1. Mathews (8) states "most tests in physical education should show re l i a b i l i t y within the range .90 - .99." He further adds, "tests objective in nature should give highly consistent results when being measured. Therefore, when evaluating tests in terms of r e l i a b i l i t y containing 40 such objective measurements, one should expect the coefficients of correlation to f a l l within this range in order to be acceptable." The analysis of the test results would appear to support a conclusion that the test is valid and reliable, and therefore, i s a useful means of obtaining an effective measure of the soccer ability of university students. The test validity and re l i a b i l i t y correlations are superior to any produced, by acceptable statistical methods, in other wall-volley-type tests, and are also higher than any obtained in battery-type tests. Thus, the test appears to be a useful, economical and a quick means of evaluating soccer s k i l l . The test lends i t s e l f to use by the coach or physical educator who wishes a means of grading and classifying large groups of subjects. The results of the sixth group tested reveal that players do improve their score in repeated trial s and the amount of improvement decreases after several tests. There appears to be a tendency for subjects to reach a peak performance score. The sixth group scores reveal a reduced range after five t r i a l s , but there i s also a reduced mean score. Thus, repeated tr i a l s may produce a general increase in performance score, but after a top performance i s achieved there appears to be a tendency for the subject to regress in performance. This may be due to reduced interest in the test after top performance. In the f i r s t t r i a l eleven subjects score in the "below average" category, while after four t r i a l s only three subjects are 41 scoring "below average." This reflects considerable general improvement. It would therefore be more appropriate to allow the subjects practice equivalent to two or three performances of the test to ensure that each subject would do his best. 42 REFERENCES 1. Mathews, Donald, K., "Measurement in Physical Education", V.B. Saunders Co., London, 1961, p.24. 2. Garrett, Henry, E., Statistics in Psychology and Education. 5th ed., Longmans, Green and Co., New York, 1959, p. 354. 3. Mathews, bp.cit.. p. 4. Schaufele, Evelyn, F., "The Establishment of Objective Tests for Girls of the Ninth and Tenth Grades to Determine Soccer Ability". (Unpublished Master's Thesis Abstract, Springfield College, 1951). 5. Mitchell, Reid, (Unpublished Master's Thesis, University of Oregon, 1963). 6. MacDonald, Lloyd, G., "The Construction of a Kicking S k i l l Test as an Index of General Soccer Ability". (Unpublished Master's Thesis Abstract, Springfield College, 1951).. 7. Garrett, op.cit.. p. 343. 8. Mathews, op.cit.„ CHAPTER VII SUMMARY, CONCLUSIONS AND RECOMMENDATIONS This study was an attempt to develop and establish the usefulness of a wall-volley-type test as a means of measuring the soccer s k i l l of players participating in soccer at The University of British Columbia,, The total group tested was 75 subjects. The total number was composed of 5 teams or groups of 15 subjects. Each group was chosen as being representative of the various levels of soccer played at the University. The performance scores were correlated against a rank order scale. The rank order scale was the external criterion used i n the test to determine test validity. The final test form was developed as a result of earlier experiments with other subjects to determine the distance of the restraining line; the number of balls to be used, and the method of scoring. The test consisted of three trials of 30-seconds each of volleying a regulation rubber soccer ball from behind a 15 foot restraining line against a target area 8 feet high by 24 feet wide. The ball was put in play by dropping i t from the hands at waist height. The subjects of each group used in the test were rank-ordered by the experimenter from one to fifteen, in order of ability. The subjects were observed playing over a twelve-game 44 period before the rank orders vere finished* The performance scores of the subjects were correlated with the experimenter's rank-ordering of subjects to determine test validity. The experimenter also used a quintiserial correlation technique to provide a total test validity coefficient. The total test validity score was .856 using the quintiserial method, while the five-group validity coefficients were: Thunderbirds Chiefs Braves P.E. Majors Required Programme .577 .840 .812 .944 .975 The experimenter f e l t that the low validity correlation found for the Thunderbird group was due mainly to conditions conforming to the old saying "the spectator sees most of the game." (The experimenter as a playing member of the Thunderbird team made his observations from a participant standpoint). The test results show that i t has the highest validity coefficient of any soccer wall-volley-type test currently mentioned in the literature. The results also show that rank-ordering, although a subjective technique, i s quite satisfactory when the observer is qualified. The r e l i a b i l i t y of the test merely confirms Garrett's (l) statement that "a highly valid test cannot be unreliable." The groups tested appeared homogeneous as regards high or low ability within groups and heterogeneous between groups. This conclusion was based on the range of scores and the size of group standard deviations. Thus, for example, the Thunderbirds were a 45 homogeneous group of high a b i l i t y . The Physical Education Majorssvere a homogeneous group of low a b i l i t y . The test would appear to serve the purpose of testing; soccer s k i l l of the University students, and provides r e l i a b l e , economic and time saving means of grading and c l a s s i f y i n g large numbers of university students. The test lends i t s e l f to the development of norms as a basis f o r c l a s s i f y i n g students i n f i v e d i s t i n c t categories of soccer a b i l i t y : superior, good, average, below average, and poor. The results of the tests of the 16 members of the 6th group would indicate that the players do improve th e i r scores with regular practice, and that such improvement decreases as the player appears to approach or reach a peak performance score. CONCLUSIONS 1. On the basis of present evidence, this wall-volley test appears to be a v a l i d and r e l i a b l e instrument for measuring soccer s k i l l . 2. Repetition of the test would permit the development of more refined norms. 3. The test r e l i a b i l i t y and v a l i d i t y indicate that i t i s a more appropriate measure of soccer a b i l i t y than any existing battery or single item type tests i l l u s t r a t e d i n the l i t e r a t u r e . 4. The test i s most useful f o r teachers and coaches who require an economic and time saving means of grading and c l a s s i f y i n g large groups of players. 5. The test may be used to categorize students as superior, good, average, below average and poor. 46 6. Repeated practice of the test does cause improvement in performance. 7. Continued practice of the test is a good means of improving soccer ability. 8. Use of the wall-volley technique is a useful device for introducing and developing soccer interest and ability i n students. RECOMMENDATIONS 1. It is suggested that test efficiency be further tested by repeating i t with further groups and greater numbers of subjects. 2. That the test scores be used as a means of grading students in one of five categories: superior, good, average, below average and poor, rather than the teacher or coach subjectively assess ability. 3. That the number of tr i a l s remain at three and the aggregate score be taken. 4. That the test be conducted indoors to determine usefulness. 5. That tr i a l s be conducted with appropriate groups to determine suitable dimensions and test conditions. 6. That tables of norms be developed for varsity, high school and elementary school students. 4 7 REFERENCES 1. Garrett, Henry, E., Statistics in Psychology and Education. 5th Ed., Longmans, Green, and Co., Inc. New York, 18, 1960, p. 361. APPENDIX 49 APPENDIX TABLE 1 Thunderbird Scores tank T r i a l s Aggregate Performance Group )rder 1 2 3 Order Data 1 18 19 21 58 1 2 16 17 15 48 7 3 17 18 21 56 2 4 17 19 18 54 3 5 16 18 13 47 8.5 Mean 48.00 6 16 15 14 45 11 Standard 7 14 16 14 44 12.5 Deviation 8 13 16 12 41 14 4.99 9 15 14 17 46 10 Rank. D i f f . 10 18 16 17 51 4 Correlation 11 16 13 21 50 5 .576 12 16 14 17 47 8.5 13 16 16 17 49 6 14 15 13 16 44 12.5 15 13 14 13 40 15 Totals 236 238 246 720 TABLE 2 Chief Scores Rank Trials Aggregate Performance Group Order 1 2 3 Order Data 1 16 15 19 50 1 2 14 12 17 43 3.5 3 15 12 12 39 6.5 4 13 14 16 43 3.5 5 13 11 11 35 9 Mean 36.93 6 14 13 12 39 6.5 7 15 8 15 38 8 S.D. = 6.72 8 14 17 15 46 2 9 15 11 15 41 5 10 10 14 10 34 10 11 11 9 10 30 12 Rank Diff. 12 13 9 9 31 11 Correlation 13 10 8 11 29 13.5 = .840 14 9 8 10 27 15 15 6 12 11 29 13.5 Totals 188 173 193 554 51 TABLE 3 Braves Rank Trials Aggregate Performance Group Order 1 2 3 Order Data 1 16 17 14 47 1 2 14 9 15 38 4 3 16 8 15 39 3 Mean 32.60 4 10 12 13 35 6 5 13 13 11 37 5 S.D.= 6.26 6 11 9 12 32 7 7 5 9 16 30 8.5 8 11 5 13 29 10.5 9 9 11 9 29 10.5 10 9 9 12 30 8.5 11 15 12 13 40 2 Rank Diff. 12 12 8 6 26 13 Correlation 13 9 6 12 27 12 .812 14 10 7 8 25 14.5 15 9 8 8 25 14.5 Totals 169 143 177 489 52 TABLE 4 -P.E. Majors Rank Trials Aggregate Performance Group Drder 1 2 J3 Order Data 1 13 9 12 34 1 2 9 11 12 32 3 3 10 9 11 30 4 4 13 11 9 33 2 5 12 8 8 28 5 Mean 25.66 6 10 8 9 27 6 7 8 8 8 24 8.5 S.D.= 4.59 8 7 8 10 25 7 9 9 9 7 24 8.5 10 8 7 6 21 13 11 7 8 6 21 13 Rank Diff. 12 8 6 9 23 10 Correlation 13 8 6 7 21 13 .944 14 7 8 6 21 13 15 9 5 7 21 13 Totals 138 120 127 385 53 TABLE 5 Required Programme Rank Trials Aggregate Performance Group Order 1 2 3 Order Data 1 12 12 12 36 1 2 13 11 9 33 2 3 10 9 11 30 3 4 10 9 8 27 5 5 7 11 10 28 4 6 8 8 8 24 7 7 7 8 10 25 6 8 9 8 6 23 9 9 8 6 7 21 11 10 9 7 7 23 9 11 8 6 9 23 9 12 8 5 7 20 12 13 5 9 4 18 13 14 5 7 4 15 14 15 5 5 5 15 15 totals 124 121 117 362 Mean 24.20 S.D.= 5.76 Rank Diff. Correlation .975 54 TABLE 6 Total Group Aggregate Scores Rank Thunderbirds Chiefs Braves P.E. Majors Read. Group Order Prog. Data 1 58 50 47 34 36 2 48 43 38 32 33 N = 75 3 56 39 39 30 30 4 54 • 43 35 33 27 Aggregate 5 47 35 37 28 28 score 6 45 39 32 27 24 = 2510 7 44 38 30 24 25 8 41 46 29 25 23 Mean 9 46 41 29 24 21 = 33.47 10 51 34 30 21 23 11 50 30 40 21 23 Median 12 47 31 26 23 20 = 29 13 49 29 27 21 18 S.D. 14 44 27 25 21 16 = 10.33 15 40 29 25 21 15 Totals 720 554 489 385 362 Mean 48.00 36.93 32.60 25.66 24.20 Rank .576 .840 .812 .944 .975 Difference Correlation 55 The Rank Difference Correlations for each group were found by using the formula: r = 1 - 6 x D 2 N(N-l) where r = coefficient of correlation form rank differences. 2 ED = the sum of the squares of the differences i n rank. N = the number of subjects or paired rankings. Thus, for the Thunderbirds the formula is replaced by: r = 1 - 6 x 237 15 x 224 = .576 Substituting in the formula for the Chiefs we have: r = 1 - 6 x 89.50 15 x 224 = .840 Substituting in the formula for the Braves we have: r = 1 - 6 x 105.50 15 x 224 = .812 Substituting in the formula for the P.E. Majors we have: r = 1 - 6 x 31.50 15 x 224 = .944 56 And substituting i n the formula f o r the Required Programme Group we have: r = 1 - 6 x 14 15 x 224 = .975 The formula f o r q u i n t i s e r i a l correlation (a correlation c o e f f i c i e n t f o r the t o t a l group) i s expressed as: r quint = ZaYa(Zb-Za) l b (Zc-Zb) Yc (Zd-Zc) Yd-Zd Ye Oy Za 2 (Zb-Za) 2 (Ze-Zb) 2 (Zd-Zc) 2 Zd 2 = .28Q0x48(.3863-.2800)36.93 (.3863)32.60 (.2800-3863)25.66-2800x24 10.34 .28002 (.3863-2800)2 (.3863-.3863)2 (.2800-.3863)2 .2800' .20 .20 .20 .20 .20 - 7.96 9.26 = .856 "The effect of s e r i a l correlation i s to normalize the segmented d i s t r i b u t i o n at the time that the correlation c o e f f i c i e n t i s obtained. I f the number of segments i s large, and i f the segmented variable i s already normally distributed, the resulting correlation w i l l be the same as a Pearson Product Moment correlation. Symbolism The following symbolism w i l l be adopted: Let y be a continuous variable, x be a continuous segmented variable, normally distributed, and r be the c o e f f i c i e n t of correlation (linear) between x and y. 5 7 Let a = the proportion of cases in the top right-most segment of the x distribution b = the proportion of cases in the second highest segment, c = the proportion of cases in the third highest segment, etc., and f = the proportion of cases in the f-th segment of the distribution. Then a b c ... = 1. Let qa = a, qb = a b, qc = a b c, etc •, qf = a b ... f = the area above the left boundary of f-th segment, and qf-1 = a b ... up to but not including f = the area above the right boundary of the f-th segment. Let Za = the ordinate of the normal curve, assuming a unit normal distribution at qa, Zb = the ordinate of the unit normal curve at qb, etc., Zf = the ordinate at qf, and Zf-1 = the ordinate at qf-1. Let l a = the mean of the y's in the top (right most) segment of the x distribution. Ib = the mean of the y's in the second highest segment, etc., and Tf = the mean of the y's in the f-th segment. Let Xa = the mean of the x's in the top segment of the x distribution, etc., and Xf = the mean of the x's in the f-th segment. 58 Except for a few modifications, conventional symbolism has been adhered to. The symbol q i s sometimes taken in the literature to represent the proportion of cases in one of the segments (i.e. 1 2 q = q - q , a decumulated frequency), as well as the cumulation of proportions or frequencies from a given line of truncation to the end of the curve. In tables of the normal probability integral oriented in terms of q, and therefore, in this paper q always represents the area from the given line of truncation to the end of the curve. Since the normal curve is symmetrical, the ordinates z are equivalent for complementary q's. Consequently most normal tables oriented in terms of q carry the argument only from zero to .500, and i t is there necessary to consult the complement of q for values of q higher than .500. This, of course, does not disturb the meaning of q. In this paper the upper segments of the x distribution (the desirable pole of the trait or measure in question) are placed to the right of the lower (or less desirable) segments of the normal curve." (l) The r e l i a b i l i t y coefficients of correlation were obtained using the product-moment formula, r = N xv - ( x )( y ) y [ N x 2 - ( x ) ^ N y 2 - ( y ) 2J where r = the coefficient of correlation of r e l i a b i l i t y between two t r i a l s x = the sum of the f i r s t t r i a l scores, y = the sum of the second t r i a l scores. 59 2 x = the sum of the squares of the f i r s t scores. 2 y = the sum of the squares of the second scores. xy = the sum of the f i r s t score times the second scores. N = the number of subjects Thus, substituting i n t r i a l s 1 and 3 we found, r 75 x 822 - (45 x 12) J[l5 x 959 - ( 4 5 ) ^ [l5 x 1126 - ( 1 2 ) ^ = .796 Substituting for t r i a l s 2 and 3 we f i n d , r = 75 x 831 - (31 x 12)  J £75 x 959 - ( 3 1 ) ^ ^ 7 5 x 1726 - (12 = .801 Substituting i n the formula for t r i a l s 1 and 2 we f i n d , r = 75 x 751 - (45 x 31)  Jfj5 x 959 - ( - 4 5 ) ^ 7 5 x 934 - ( 3 1 ) ^ = .742 Applying the Spearman Brown Prophecy formula for extending the correlation between n forms of a test and n comparable forms, we have, = m l+(n-l)^„ where- = the correlation between n forms of the test and n alternate forms 60 •fif = the re l i a b i l i t y coefficient of Trial; 1. Thus, by tripling the test, the Spearman Brown Prophecy formula would alter the correlations of re l i a b i l i t y as follows, Trials 1 and 3 r3111 3 x .796 1 +(2 x .796) = .900 for tria l s 2 and 3 7^3111 = 3 x .801 1 + (2 x .801) = .924 for trials 1 and 2 / S i l l = 3 x .796 the total group was found by using the following formula 1 +(2 x .796) = .921 The standard deviation for the five groups and for Substituting in Thunderbird Group: = 4.99 61 Substituting in Chiefs Group: S.D. = / 21134 - (35.6)2 = 6.26 15 Substituting in Braves Group: S.D. = / 16529 - (32.60)2 = 6.34 y 15 Substituting in P.E. Major Group: S.D. = /10193 - (25.66)2 = 4.59 / 15 Substituting in Required Programme Group: S.D. = /9232 - (24.13)2 = 5.76 Substituting in total group: S.D. =/ 92022 - (33.47)2 =10.33 ,/ 75 Standard Deviation Scores for Sixth Group Trials 1st Trial S.D. J Ex 2 - (M)2 J N I 2 T 2nd Trial S.D. = / Ex - (M) J N =* _Exf. - (M)2 J N 4th Trial S.D. =f _Ex2_ - (M)2 = / 20256 - (36)' = 7.39 N /11?1? - (26.44)2 ' 16 /l5453 - (29.94)2 / 1 6 / 17815 - (33.27)2 15 ' 20256 - 2 15 5th Trial S.D. =/ Ex 2 - (M)2 = / 15909 - (34.38)2 = 6.46 N J 13 62 Smoothing Frequency Distribution Curve Mid-Point of Interval 13.5 16.5 19.5 22.5 25.5 28.5 Frequency 0 10 10 Correction Calculation 0 + 0 + 2 0 + 2 + 2 3 2 + 2 + 1 0 3 2 + 1 0 + 8 3 1 0 + 8 + 1 0 8 + 1 0 + 8 Correction .66 1.33 4.66 6.66 9.33 8.66 31.5 34.5 1 0 + 8 + 6 3 8 + 6 + 4 8.00 6.00 37.5 40.5 43.5 46.5 49.5 52.5 6 + 4 + 7 3 4 + 7 + 4 3 7 + 4 + 6 3 4 + 6 + 4 6 + 4 + 1 3 4 + 1 + 2 5.66 5.00 5.66 4.66 3.66 2.33 55.5 58.5 61.5 1 + 2 + 1 3 2 + 1 + 0 1 + 0 + 0 3 1.33 1.00 .33 63 Calculation of Percentiles P / 0 o = 57 f 75-74 x 3 = 60 1 P v = 514-71.25-71 x 3 =51.75 1 Pq0 = 48+ 67.5-67 x 3 = 48.17 4 *V = 4 5 + 63.75-61 x 3 = 46.37 6 P g 0 = 42-4-60-57 x 3 = 43.8 5 P 7 5 = 39+ 56.25-50 x 3 = 41.68 7 P;0 = 39 + 52.5-50 x 3 = 40.25 6 Ptf = 36+48.75-46 x 3 = 38.06 4 Pio = 33 + 45-40 x 3 = 35.14 7 P S 5 = 334-41.25-40 x 3 = 33.62 6 P ? 0 = 30+ 37.3-32 x 3 = 32.75 8 Vus = 30-4-33.75-32 x 3 = 30.66 8 P40 = 27+- 30.22 x 3 = 29.4 10 PJ5 = 27+ 26.25-22 x 3 = 28.27 Pjo = 27 ^ 22.5-22 x 3 =27.13 10 VlS = 24+. 18.75-14 x 3 = 25.78 8 Yxo = 24+15-14 x 3 = 24.35 8 21^11.25-4 x 3 = 23.90 10 21^7.5-4 x 3 = 22.05 10 18 -+ 3.75-2 x 3 = 20.62 2 14.9 64 a 65 REFERENCES 1. Jaspen, Nathan, "Serial Correlation", Psvchometrika, March, 1946, pp. 23-30. 'i BIBLIOGRAPHY BOOKS Caswell, John, E., Soccer for High Schools. A.S. Barnes and Co., New York, 1933. Csaknady, Jeno, Learn to Play the Hungarian Way. Hungarian Sport Publishing House, Budapest, 1954, pp. 17-71. Finney, Tom, Instructions to Young Footballers. Museum Press Limited, London, 1955, pp. 16-34. Garrett, Henry, E., Statistics in Psychology and Education. 5th ed., Longmans, Green and Co., New York, 1958, pp. 471. Guilford, J. P., Fundamental Statistics in Psychology and Education. 2nd ed., McGraw H i l l Book Co., New York, 1950, pp. 51-54. Mathews, Donald, K., Measurement in Physical Education. W. B. Saunders Co., Philadelphia, 1958, 349pp. Meisl, Willy, Soccer Revolution. Phoenix Sports Club, London, 1955, pp. 20-34. Voltmer, Edward, F., and Esslinger, Arthur, A., The Organization and Administration of Physical Education. 3rd ed., Appleton-Century-Crofts, Inc., New York, 1958, 547 pp. Winterbottom, Walter, Soccer Coaching. Naldrett Press Ltd., London, 1955. PERIODICALS B i l l C-131, An Act to encourage Fitness and Amateur Sport. As passed by the House of Commons, 25th September, 1961. Queen's Printers, Ottawa, 1961. Brady, George, F., "Preliminary Investigations of Volleyball Playing Ability". Research Quarterly, vol. 18, no. 1, March, 1945. Cornish, Clifford, "A Study of Measurement of Ability in Handball". Research Quarterly, vol. 20, no. 2, May, 1949. Dyer, Joanna, T., "Revision of Backboard Test of Tennis Ability". Research Quarterly, vol. 9, no. 1, March, 1938, p. 27. French, Esther and Cooper, Bernice, "Achievement Tests in Volleyball for High School Girls". Research Quarterly, vol. 8, no. 2, May, 1937. 67 Heath, Marjorie, L. and Rogers, Elizabeth, G., "A Study in the Use of Knowledge and S k i l l Tests in Soccer", Research  Quarterly, vol. 33, ho. 3, December, 1932. Jaspen, Nathan, "Serial Correlation", Psychometrika. March, 1946, pp. 23-30. Miller, Prances, A., "A Badminton Wall Volley Test". Research Quarterly, vol. 22, no. 2, May, 1951. Russell, Naomi, and Lange, Elizabeth, "Achievement Tests in Volleyball for Junior High School Girls", Research Quarterly. vol. 2, no. 4, December, 1940. Vanderhoof, Mildred, "Soccer S k i l l Tests", Journal of Health and  Physical Education, vol. 42, no. 3, October, 1932. Weiss, Raymond, A., and Scott, Gladys, M., "Construction of Tests", Research Methods. 2nd ed., American Association for Health, Physical Education and Recreation, Washington, D.C, 1959, pp. 239-240. THESES Bontz, Jean, "An Experiment in the Construction of a Test for Measuring Ability in Some of the Fundamental Skills Used by Fifth and Sixth Grade Children in Soccer". (Unpublished Master's Thesis, State University of Iowa, 1942). Crawford, Elinor, A., "The Development of S k i l l Test Batteries for Evaluating the Ability of Women Physical Education Major Students in Soccer and Speedball". (Unpublished Doctoral Thesis, University of Oregon, 1958). Konstantinov, J., "Construction of a Soccer Kicking Test". (Unpublished Master's Thesis, Springfield College, 1950). MacDonald, Lloyd, G., "The Construction of a Kicking S k i l l Test as an Index of General Soccer Ability". (Unpublished Master's Thesis, Springfield College, 1951). Mitchell, Reid, "A Wall-Volley Test for Measuring Soccer Ability i n Fifth and Sixth Grade Boys". (Unpublished Master's Thesis, University of Oregon, 1963). Schaufele, Evelyn, F., "The Establishment of Objective Tests for Girls of the Ninth and Tenth Grades to Determine Soccer Ability". (Unpublished Master's Thesis, State University of Iowa, 1940). 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0077413/manifest

Comment

Related Items