This article was downloaded by: [University of Toronto Libraries] On: 30 January 2015, At: 11:20 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Sports Sciences Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/rjsp20
Withdrawal predictors among physical and performance characteristics of female competitive gymnasts a
b
Koenraad J. Lindner , Dennis J. Caine & David P. Johns
a
a
Faculty of Physical Education and Recreation Studies , University of Manitoba , Winnipeg, Manitoba, R3T2N2, Canada b
Faculty of Education , University of New Brunswick , Fredericton, New Brunswick, E3B 6E3, Canada Published online: 14 Nov 2007.
To cite this article: Koenraad J. Lindner , Dennis J. Caine & David P. Johns (1991) Withdrawal predictors among physical and performance characteristics of female competitive gymnasts, Journal of Sports Sciences, 9:3, 259-272, DOI: 10.1080/02640419108729888 To link to this article: http://dx.doi.org/10.1080/02640419108729888
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sublicensing, systematic supply, or distribution in any form to anyone is expressly
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Sports Sciences, 1991, 9, 259-272
Withdrawal predictors among physical and performance characteristics of female competitive gymnasts Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
KOENRAAD J. LINDNER,1 DENNIS J. CAINE2 and DAVID P. JOHNS 1 1 Faculty of Physical Education and Recreation Studies, University of Manitoba, Winnipeg, Manitoba R3T2N2 and 2Faculty of Education, University of New Brunswick, Fredericton, New Brunswick E3B 6E3, Canada
Accepted 26 May 1990
Abstract Future drop-outs (n = 27) and continuing (n=41) female competitive gymnasts were compared with respect to their physical, performance and injury characteristics measured through a large battery of tests completed while they were participants in competition. Included were anthropometric variables, body composition and somatotype, strength, flexibility, endurance, power, speed and balance measurements, and previous injury information. Differences were examined through t-tests and through discriminant analyses of principal components. The results indicate that the drop-outs were as a group distinguishable from the continuing gymnasts: they were significantly older, taller and heavier which may account for the finding that they were significantly stronger, more powerful and faster, and had greater endurance. They also had a slightly more linear/ectomorphic physique with less muscularity, and performed better on most flexibility tests. Only in a gymnastic-specific flexibility variable did the continuing gymnasts out-perform the drop-outs. It was concluded that it was mainly the age factor, and presumably the social and psychological factors associated with it, that distinguished the future drop-outs from the continuing gymnasts, and that factors related to the physical make-up and performance capacities of gymnasts cannot readily predict withdrawal from the sport. Keywords: Drop-outs, gymnastics, characteristics, physical performance, injury.
Introduction The extensive literature on attrition in youth sport, recently reviewed by Weiss and Petlichkoff (1989), has primarily focused on the former athletes' superficial reasons for leaving the sport, and on the social and psychological factors underlying these reasons. In the first line of study, drop-outs are typically asked in a questionnaire to indicate the extent of their agreement with a number of stated withdrawal reasons (e.g. Johns et al., 1990; Seye and Salmela, 1987; Klint and Weiss, 1986; Narciso et al., 1984), while more recently researchers have started to formulate theories on sport withdrawal to examine the factors underlying these superficial reasons. Such theories include Competence Motivation (Weiss, 1986; Burton and Martens, 1986), Burn-out (Fender, 1989; Feigley, 1987; Smith, 1986a), 0264-0414/91 $03.00 + .12 © 1991 E. & F.N. Spon
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
260
Lindner et al.
Social Exchange (Gould, 1987; Smith, 1986b) and Alternative Status Cultures (Johns, 1980; Johns et al, 1990). Lindner et al. (1991) have proposed that, while these social and psychological factors may account for a proportion of the withdrawal cases, many others seem to be related to natural developmental change. The drop-out, who is typically between 10 and 17 years of age (Gould and Horn, 1984), undergoes great physical and psychological changes, which may well influence the athlete's decision to withdraw from the sport. While the socio-physiological aspects have received some attention, relatively little study has been done of the physical and performance characteristics of drop-outs as compared to those of continuing athletes. Such research would be important, for instance, to verify whether perceived lack of competence in drop-outs (as proposed in the Competence Motivation Theory) corresponds with actual inferior performance levels, or to examine whether physical characteristics and changes therein are different for drop-outs and continuing athletes. Research pertaining to the anthropometric characteristics of athletes is complicated by the varying growth and development and the participation categories of the subjects studied. Nadgir (1988) and Broekhoff et al. (1986) employed a 'phantom' strategy, where the bodily dimensions are scaled to a standardized stature of a unisex human model, the phantom, and expressed in standard scores. These studies have confirmed previous findings based on absolute values, such as larger upper-arm girth and humerus width, broader shoulders and narrower hips for female gymnasts, as well as their being lighter than age-matched controls (Vercruissen, 1984; Bernink et al., 1983; Brisson et al., 1983; Nelson et al., 1983; Beunen et al., 1981). These findings suggest a performance advantage in gymnastic moves requiring support and suspension. The anthropometric profile of the successful female gymnast thus appears to include a relative small stature, light weight, narrow hips and well-developed upper extremities. The somatotype of gymnasts has been almost without exception reported to be ectomesomorphic, i.e. mesomorphy as the dominant component and ectomorphy greater than endomorphy (Caine et al., 1989; Nadgir, 1988; Carter and Brallier, 1988; Broekhoff et al., 1986; Carter, 1981). It is evident that a physique model exists which is characteristic of the successful gymnast (Salmela, 1987). The biomechanical aspects of women's gymnastics appear to favour a linear physique with high mesomorphy and low endomorphy with a resultant increased strength-weight ratio (Bale and Goodway, 1987; Salmela, 1979). There has been considerable interest in the literature for gymnasts' body composition, determined through hydrostatic weighing (Nadgir, 1988; Bale and Goodway, 1987; Eston and Maridaki, 1986; Broekhoff et al., 1986; Vercruissen, 1984; Slaughter et al, 1981; Thorland et al., 1981, Sinning, 1978), or through derivation from anthropometric measurements (Theintz et al, 1989; Lindner and Caine, 1989; Sands et al, 1987; Bale and Goodway, 1987; Haywood et al, 1986; Caldarone et al, 1986; Zonderland et al, 1985; Brisson et al, 1983; Falls and Humphrey, 1978; Novak et al, 1977). The former method tends to yield lower body fat estimates than the latter (Bale and Goodway, 1987; Fleck, 1983). Body fatness and density typically increase in absolute values with age in the general population, but in gymnasts this developmental change is less clear (Haywood et al, 1986; Peltenburg et al, 1984). When gymnasts are compared to non-athletic controls, their body fat percentage is typically significantly lower (Theintz et al, 1989; Broekhoff et al, 1986; Eston and Maridaki, 1986). Although a substantial number of studies have measured the motor performance characteristics of female gymnasts (Sands et al, 1987; Bale and Goodway, 1987; Haywood et al, 1986; Nelson et al, 1983; Kirby et al, 1981; Beunen et al, 1981), none has involved an
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
Withdrawal predictors in gymnastics
261
extensive performance test battery. Bajin (1987) has stressed the importance of physical ability to gymnastic success and several authors have reported better performance in tests of strength and flexibility by female gymnasts than their non-athletic peers (Beunen et al., 1981; Kirby et al., 1981) and than peers involved in other sports (e.g. Haywood et al., 1986). The above review indicates that the gymnast has physical and performance characteristics that distinguishes her from her non-athletic peer, and there appears to be some consensus about the qualities a gymnast must possess to be successful. Differences in these characteristics between drop-outs from gymnastics and those who persist in their sport have not been investigated to date. In the present study, competitive gymnasts were given a large battery of physical measurements and performance tests at the start of a 3-year surveillance period. The purpose of the present analysis was to compare the physical and performance characteristics of those gymnasts who would drop out of the sport during the surveillance period with those who would continue their competitive participation. Such comparison served to identify any characteristics that might be indicative of dropping out in the future.
Method Subjects The sample consisted of 68 female competitive gymnasts from three prominent private clubs which instruct gymnasts from beginners' to international levels. They were volunteers from a group of 178 competitive gymnasts who were subjects in a 3-year prospective epidemiological injury investigation (Lindner and Caine, 1988, 1990). Only this sub-sample was given the battery of anthropometric and performance tests. The subjects were grouped according to age and competitive level. The age range was divided into three intervals: below 10.5 years, designated as '9' (n = 27, mean age = 9.69); between 10.6 and 12.5 years, designated a s ' 11' (n = 21, mean age =11.49); and above 12.5 years, designated as '13' (n = 20, mean age= 13.60). Two levels of competition were selected: Canadian Provincial levels 1,2, and 3 combined, designated as 'low' (n = 43, mean age = 10.73), and Provincial level 4 and National Elite level combined, designated as 'high' (n = 25, mean age =12.54). Of this sample, 41 were continuing gymnasts, while 27 withdrew from the sport at some time during the surveillance period. The latter gymnasts are referred to as 'drop-outs' and they left their sport on average a little over 1 year after the testing (x= 13.33 months, S.D. = 6.25, range=4-28). The withdrawal times for the three groups were as follows: age group 9 (x=12.8 months, S.D. = 6.18), age group 11 (x = 13.7, S.D. = 5.4) and age group 13 (x = 13.3, S.D. = 7.0).
Tests and variables The gymnasts were given a battery of anthropometric and performance tests in the first year of the study. Eighty tests and measurements yielded - with the derived measurements and categorical variables - a total of 100 variables. Of the anthropometric variables, 7 were height and weight measurements (see Table 1), 14 were body segmental lengths, 14 were circumferences, 10 were breadth measurements, and 12 were skinfold measurements and body composition estimates, i.e. % body fat (from the body density formula suggested by
262
Lindner et al.
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
Jackson et al., 1980, and Siri's conversion formula as reported in Johnson and Nelson, 1986), O-scale adiposity rating (Ross et al., 1987) and somatotype (Carter, 1980). The performance variables included: • 7 strength measurements, i.e. 5 cable tensiometer tests (Clarke and Munroe, 1970) and 2 grip strengths; • 17 flexibility measures, i.e. 9 flexometer tests (Leighton, 1983), the sit-reach test (Scott and French, 1959) and a set of 7 field tests for gymnasts developed by Bajin (1978); • 3 endurance measures (bar dips, chin-ups, leg lifts); • 2 power measurements, i.e. the standing broad jump and the vertical jump (Mathews and Fox, 1976); • 2 speed measurements, i.e. a 20-m run and simple total body reaction time; and • 1 balance test on the stabilometer.
Previous injuries, i.e. the total number of injuries sustained prior to the surveillance period as reported by the gymnasts in a background information questionnaire, was also included in the comparisons between the drop-outs and the continuing gymnasts. Data analysis The means and standard deviations for the two categories of gymnasts were calculated and the differences tested through t-tests. The number of variables was then reduced (a) by eliminating the base variables for any kept composite variables (e.g. the six skinfold measurements) or deleting the composite or derived variables for any kept base variables (e.g. ectomorphic somatotype component), (b) by selecting the variables with the highest eigenvector values from clear groupings of variables by separate principal component analyses of the anthropometric and strength and flexibility variables (e.g. thigh circumference representing the four extremity circumferences), and (c) by selecting one of the groups of variables measuring approximately the same ability (e.g. standing broad jump for power). Further multivariate reduction was then accomplished through principal component analyses (SAS Princomp Procedure) of the 22 selected anthropometric and performance variables. The resulting components are listed in Table 2. Component 1 was a large general factor representing body size, strength, power and speed. The second component was a combination of endurance and flexibility. The body fat component was identifiable from the skinfolds and the ponderal index variables, but also contained a flexibility element. The flexometer and sit-reach tests made up the bulk of the fourth component, and because the gymnastic-specific items were not included, this component was labelled 'extent flexibility'. The fifth principal component was a combination of specific flexibility, balance, speed and arm endurance, while component 6 combined a gymnastic-specific flexibility item with strength, power and balance. The seventh and eighth principal components had speed and balance as main elements respectively. These principal components were then used as variables in discriminant analyses along with the previous injury variable. Theses analyses (SAS Discrim Procedure) served to answer the following question: How distinguishable are continuing gymnasts from those who would drop out during the study on the basis of the principal components and the injury variable? Subsequent stepwise discriminant analyses (SAS Stepdisc Procedure) indicated which of the
Withdrawal predictors in gymnastics
263
Table 1. The variables in the test battery and in the principal component analysis
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
Age Height Weight Sit height Ponderal index" Quetelet index Lengths Leg length Left total arm" Right total arm" Left upper arm Left lower arm Left hand Left total leg" Right total leg Left lower leg Right lower leg Left upper leg Right upper leg Left lower total Right lower total Left foot Widths Biacromial" Bideltoid Bicristal" Bitrochanteric Chest Biepicondylar fern. Ankle Biepicondylar hum. Wrist Hand
Circumferences Ankle Calf Forearm Thigh" Hip Lower abdomen Middle abdomen Upper abdomen" Upper arm, flexed Upper arm, extended Chest Knee Wrist Neck Skinfolds Subscapular Triceps Abdominal Suprailiac Thigh Calf Sum 6 skinfolds" % Body fat O-scale adiposity Cable tensiometer Shoulder flexion Shoulder adduction" Knee extension" Knee flexion Ankle plantar flexion Previous injuries Vital capacity
Somatotype Flexometer Shoulder flexion Shoulder abduction Elbow flex.-ext. Radius-ulnar flex.-ext." Knee flex.-ext. Ankle flexion Trunk flexion Trunk extension" Wrist flex.-ext. Splits Forward right" Forward left Sideways Leg raises Forward left Forward right" Sideways left Sideways right Motor performance Grip strength, left Grip strength, right Sit-reach" 20-m run" Leg lifts" Standing broad jump" Bar dips" Chin-ups" Vertical jump Simple reaction time" Balance"
"These variables were selected for further multivariate analyses. components were most important in discriminating between the future drop-outs and the continuing gymnasts.
Results Anthropometry, body composition and somatotype
The future drop-outs as a group were significantly older, taller and heavier than the continuing gymnasts at the time of the testing (Table 3). This difference remained when the gymnasts were categorized according to competitive level, but largely disappeared within the
264
Lindner et al.
Table 2. Principal components for the set of 22 anthropometric and performance variables and their eigenvectors" Component'
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
%var.' Ponderal index Arm length Leg length Biacromial width Bicristal width Thigh circumf. Upper abdomen circumf. Skinfolds Shoulder adduct. Knee extension Radius-ulnar Trunk extension Spilts, forward right Leg raises, forward right Sit-reach 20-mrun Leg lifts
Standing broad jump Bar dips Chin-ups Simple reaction time Balance
Cl
C2
C3
C4
C5
C6
C7
C8
31.05
16.66
8.67
7.22
5.59
4.74
4.63
3.47
-0.59
0.24 0.23
0.32 0.33 0.27 0.32 0.33
-0.31 -0.22
0.26 0.25
0.28
0.30
-0.20 -0.27
0.30 0.29 flex.
0.51 -0.28
0.30
0.29 0.28
0.41 0.36
0.31 0.39 0.27
0.23 0.52
-0.29
-0.23
0.47
-0.27
0.42
-0.25
0.36 -0.40
0.22
0.42
0.32 0.35
-0.31
0.30
-0.25
0.43
0.30 -0.36 -0.31 -0.22 0.24
-0.38 0.85 0.55
"Only variables with loadings of over 0.20 are listed. *% var., percentage of total variation accounted for. C C1, body size, strength, power and speed; C2, endurance and flexibility; C3, body fat and flexibility; C4, extent flexibility; C5, balance, speed and arm flexibility and strength; C6, gymnastic-specific flexibility and power vs endurance; C7, speed; C8, balance. age intervals. The drop-outs also had broader shoulders in all groups, significantly so in the whole sample and the two competitive levels. There were no significant differences in body fat measurements for any of the groups and no clear trend was apparent. The drop-outs tended to score lower on the mesomorphy somatotype component, with the exception of the two younger age groups, but only for age group 13 was the comparison statistically significant. The mean somatotype for the whole sample was 1.95 (endo), 4.18 (meso) and 3.29 (ecto), while the comparison between the dropouts and the continuing gymnasts yielded somatotypes of 1.89-4.03-3.44 and 2.04-4.29-3.22 respectively. The means of the previous injuries variable tended to be higher (but not significantly so) for the drop-outs in all comparisons except in the age 9 interval.
Table 3. Means (x) and standard deviations (S.D.) of selected anthroporaetric, somatotype and body composition variables for drop-outs (D) and continuing gymnasts (C) All
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
Variable
Low-level
High-level
D (" = 27)
C (n=41)
D (n = 13)
C (n = 30)
D (n=14)
C (n = l l )
12.6* 1.7
10.6 1.4
12.1** 1.5
10.2 1.2
13.1 1.7
11.9 1.3
147.5* 10.0
136.5 9.6
146.9** 11.4
135.2 9.3
148.1ft 7.7
39.0* 6.9
31.7 6.7
37.4** 6.6
30.8 6.6
292.2* 21.9
259.6 26.1
291.9* 21.9
451.1* 34.6
416.4 38.9
48.4 9.2
Age 9 group D (n = 4)
C ("=23:
Age 11 group D )
Age 13 group
C ( i = i 4 ;1
D (n = 16)
C (n = 4)
Age (years) X
S.D.
0.7
9.6 0.8
11.2 0.6
11.6 0.5
13.7 1.1
13.2 0.7
139.8 10.2
134.0 10.9
130.6 6.3
142.9 6.3
142.7 7.8
152.9 5.2
148.4 6.7
40.6ft 7.0
34.0 6.6
30.3 5.4
27.9 3.9
34.9 6.1
34.8 4.7
42.6 4.8
42.5 8.7
253.8 26.7
292.5ft 22.7
275.4 17.0
263.5 16.8
250.4 22.6
284.0 24.3
264.6 24.5
302.9 12.8
295.5 14.9
437.9 31.1
412.5 40.7
463.3f 34.2
427.0 33.1
423.8 37.4
396.2 28.5
437.7 38.0
434.6 25.7
463.8 27.7
469.3 55.0
50.0 15.7
48.9 10.6
52.7 16.7
47.9 8.1
42.4 9.5
50.3 16.2
48.8 14.5
49.7 8.3
52.1 16.9
47.3 8.0
49.4 22.1
11.8 2.4
11.9 3.3
12.4 3.1
12.4 3.4
11.3 1.5
10.4 2.2
13.3 5.2
11.7 3.1
11.7 1.8
12.3 3.5
11.5 1.7
11.0 3.8
1.9 0.6
2.0 0.9
1.9 0.7
2.2 1.0
1.9 0.5
1.7 0.5
2.1 1.3
2.0 0.8
1.9 0.3
2.1 1.1
1.9 0.5
2.0 0.7
4.0 0.8
4.3 0.8
3.9 0.9
4.2 0.7
4.1 0.6
4.6 0.6
4.5 1.1
4.4 0.6
4.3 0.8
4.0 0.7
0.7
3.8ft
4.8 1.0
3.4 0.9
3.2 0.9
3.7 1.0
3.2 1.0
3.2 0.8
3.2 0.5
3.0 1.1
3.2 0.8
3.2 1.2
3.5 1.0
3.7 0.8
2.8 1.3
1.44 1.12
1.00 1.16
1.08 0.76
0.73 0.99
1.79 1.31
1.73 1.35
0.50 0.52
0.67 0.93
1.57 1.27
1.36 1.22
1.63 1.09
1.50 1.92
10.4ft
Height (cm) X
S.D.
.
1o* 1
Weight (kg) X
S.D.
8
Biac. width (mm)" X
S.D.
Thigh circumf. (mm) X
S.D.
Skinfolds* X
S.D.
% Fat X
S.D.
Endomorph.
c
X
S.D.
Mesomorph.11 X
S.D.
Ectomorph.6 X
S.D.
Prev. inj. (no./ X
S.D.
"Biac. width, biacromial width; 'skinfolds, sum of six skinfolds; "endomorph., endomorphic somatotype component; ''mesomorph., mesomorphic somatotype component; "ectomorph., ectomorphic somatotype component; •'prev. inj., previous injuries. *P
Journal of Sports Sciences Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/rjsp20
Withdrawal predictors among physical and performance characteristics of female competitive gymnasts a
b
Koenraad J. Lindner , Dennis J. Caine & David P. Johns
a
a
Faculty of Physical Education and Recreation Studies , University of Manitoba , Winnipeg, Manitoba, R3T2N2, Canada b
Faculty of Education , University of New Brunswick , Fredericton, New Brunswick, E3B 6E3, Canada Published online: 14 Nov 2007.
To cite this article: Koenraad J. Lindner , Dennis J. Caine & David P. Johns (1991) Withdrawal predictors among physical and performance characteristics of female competitive gymnasts, Journal of Sports Sciences, 9:3, 259-272, DOI: 10.1080/02640419108729888 To link to this article: http://dx.doi.org/10.1080/02640419108729888
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sublicensing, systematic supply, or distribution in any form to anyone is expressly
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Sports Sciences, 1991, 9, 259-272
Withdrawal predictors among physical and performance characteristics of female competitive gymnasts Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
KOENRAAD J. LINDNER,1 DENNIS J. CAINE2 and DAVID P. JOHNS 1 1 Faculty of Physical Education and Recreation Studies, University of Manitoba, Winnipeg, Manitoba R3T2N2 and 2Faculty of Education, University of New Brunswick, Fredericton, New Brunswick E3B 6E3, Canada
Accepted 26 May 1990
Abstract Future drop-outs (n = 27) and continuing (n=41) female competitive gymnasts were compared with respect to their physical, performance and injury characteristics measured through a large battery of tests completed while they were participants in competition. Included were anthropometric variables, body composition and somatotype, strength, flexibility, endurance, power, speed and balance measurements, and previous injury information. Differences were examined through t-tests and through discriminant analyses of principal components. The results indicate that the drop-outs were as a group distinguishable from the continuing gymnasts: they were significantly older, taller and heavier which may account for the finding that they were significantly stronger, more powerful and faster, and had greater endurance. They also had a slightly more linear/ectomorphic physique with less muscularity, and performed better on most flexibility tests. Only in a gymnastic-specific flexibility variable did the continuing gymnasts out-perform the drop-outs. It was concluded that it was mainly the age factor, and presumably the social and psychological factors associated with it, that distinguished the future drop-outs from the continuing gymnasts, and that factors related to the physical make-up and performance capacities of gymnasts cannot readily predict withdrawal from the sport. Keywords: Drop-outs, gymnastics, characteristics, physical performance, injury.
Introduction The extensive literature on attrition in youth sport, recently reviewed by Weiss and Petlichkoff (1989), has primarily focused on the former athletes' superficial reasons for leaving the sport, and on the social and psychological factors underlying these reasons. In the first line of study, drop-outs are typically asked in a questionnaire to indicate the extent of their agreement with a number of stated withdrawal reasons (e.g. Johns et al., 1990; Seye and Salmela, 1987; Klint and Weiss, 1986; Narciso et al., 1984), while more recently researchers have started to formulate theories on sport withdrawal to examine the factors underlying these superficial reasons. Such theories include Competence Motivation (Weiss, 1986; Burton and Martens, 1986), Burn-out (Fender, 1989; Feigley, 1987; Smith, 1986a), 0264-0414/91 $03.00 + .12 © 1991 E. & F.N. Spon
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
260
Lindner et al.
Social Exchange (Gould, 1987; Smith, 1986b) and Alternative Status Cultures (Johns, 1980; Johns et al, 1990). Lindner et al. (1991) have proposed that, while these social and psychological factors may account for a proportion of the withdrawal cases, many others seem to be related to natural developmental change. The drop-out, who is typically between 10 and 17 years of age (Gould and Horn, 1984), undergoes great physical and psychological changes, which may well influence the athlete's decision to withdraw from the sport. While the socio-physiological aspects have received some attention, relatively little study has been done of the physical and performance characteristics of drop-outs as compared to those of continuing athletes. Such research would be important, for instance, to verify whether perceived lack of competence in drop-outs (as proposed in the Competence Motivation Theory) corresponds with actual inferior performance levels, or to examine whether physical characteristics and changes therein are different for drop-outs and continuing athletes. Research pertaining to the anthropometric characteristics of athletes is complicated by the varying growth and development and the participation categories of the subjects studied. Nadgir (1988) and Broekhoff et al. (1986) employed a 'phantom' strategy, where the bodily dimensions are scaled to a standardized stature of a unisex human model, the phantom, and expressed in standard scores. These studies have confirmed previous findings based on absolute values, such as larger upper-arm girth and humerus width, broader shoulders and narrower hips for female gymnasts, as well as their being lighter than age-matched controls (Vercruissen, 1984; Bernink et al., 1983; Brisson et al., 1983; Nelson et al., 1983; Beunen et al., 1981). These findings suggest a performance advantage in gymnastic moves requiring support and suspension. The anthropometric profile of the successful female gymnast thus appears to include a relative small stature, light weight, narrow hips and well-developed upper extremities. The somatotype of gymnasts has been almost without exception reported to be ectomesomorphic, i.e. mesomorphy as the dominant component and ectomorphy greater than endomorphy (Caine et al., 1989; Nadgir, 1988; Carter and Brallier, 1988; Broekhoff et al., 1986; Carter, 1981). It is evident that a physique model exists which is characteristic of the successful gymnast (Salmela, 1987). The biomechanical aspects of women's gymnastics appear to favour a linear physique with high mesomorphy and low endomorphy with a resultant increased strength-weight ratio (Bale and Goodway, 1987; Salmela, 1979). There has been considerable interest in the literature for gymnasts' body composition, determined through hydrostatic weighing (Nadgir, 1988; Bale and Goodway, 1987; Eston and Maridaki, 1986; Broekhoff et al., 1986; Vercruissen, 1984; Slaughter et al, 1981; Thorland et al., 1981, Sinning, 1978), or through derivation from anthropometric measurements (Theintz et al, 1989; Lindner and Caine, 1989; Sands et al, 1987; Bale and Goodway, 1987; Haywood et al, 1986; Caldarone et al, 1986; Zonderland et al, 1985; Brisson et al, 1983; Falls and Humphrey, 1978; Novak et al, 1977). The former method tends to yield lower body fat estimates than the latter (Bale and Goodway, 1987; Fleck, 1983). Body fatness and density typically increase in absolute values with age in the general population, but in gymnasts this developmental change is less clear (Haywood et al, 1986; Peltenburg et al, 1984). When gymnasts are compared to non-athletic controls, their body fat percentage is typically significantly lower (Theintz et al, 1989; Broekhoff et al, 1986; Eston and Maridaki, 1986). Although a substantial number of studies have measured the motor performance characteristics of female gymnasts (Sands et al, 1987; Bale and Goodway, 1987; Haywood et al, 1986; Nelson et al, 1983; Kirby et al, 1981; Beunen et al, 1981), none has involved an
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
Withdrawal predictors in gymnastics
261
extensive performance test battery. Bajin (1987) has stressed the importance of physical ability to gymnastic success and several authors have reported better performance in tests of strength and flexibility by female gymnasts than their non-athletic peers (Beunen et al., 1981; Kirby et al., 1981) and than peers involved in other sports (e.g. Haywood et al., 1986). The above review indicates that the gymnast has physical and performance characteristics that distinguishes her from her non-athletic peer, and there appears to be some consensus about the qualities a gymnast must possess to be successful. Differences in these characteristics between drop-outs from gymnastics and those who persist in their sport have not been investigated to date. In the present study, competitive gymnasts were given a large battery of physical measurements and performance tests at the start of a 3-year surveillance period. The purpose of the present analysis was to compare the physical and performance characteristics of those gymnasts who would drop out of the sport during the surveillance period with those who would continue their competitive participation. Such comparison served to identify any characteristics that might be indicative of dropping out in the future.
Method Subjects The sample consisted of 68 female competitive gymnasts from three prominent private clubs which instruct gymnasts from beginners' to international levels. They were volunteers from a group of 178 competitive gymnasts who were subjects in a 3-year prospective epidemiological injury investigation (Lindner and Caine, 1988, 1990). Only this sub-sample was given the battery of anthropometric and performance tests. The subjects were grouped according to age and competitive level. The age range was divided into three intervals: below 10.5 years, designated as '9' (n = 27, mean age = 9.69); between 10.6 and 12.5 years, designated a s ' 11' (n = 21, mean age =11.49); and above 12.5 years, designated as '13' (n = 20, mean age= 13.60). Two levels of competition were selected: Canadian Provincial levels 1,2, and 3 combined, designated as 'low' (n = 43, mean age = 10.73), and Provincial level 4 and National Elite level combined, designated as 'high' (n = 25, mean age =12.54). Of this sample, 41 were continuing gymnasts, while 27 withdrew from the sport at some time during the surveillance period. The latter gymnasts are referred to as 'drop-outs' and they left their sport on average a little over 1 year after the testing (x= 13.33 months, S.D. = 6.25, range=4-28). The withdrawal times for the three groups were as follows: age group 9 (x=12.8 months, S.D. = 6.18), age group 11 (x = 13.7, S.D. = 5.4) and age group 13 (x = 13.3, S.D. = 7.0).
Tests and variables The gymnasts were given a battery of anthropometric and performance tests in the first year of the study. Eighty tests and measurements yielded - with the derived measurements and categorical variables - a total of 100 variables. Of the anthropometric variables, 7 were height and weight measurements (see Table 1), 14 were body segmental lengths, 14 were circumferences, 10 were breadth measurements, and 12 were skinfold measurements and body composition estimates, i.e. % body fat (from the body density formula suggested by
262
Lindner et al.
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
Jackson et al., 1980, and Siri's conversion formula as reported in Johnson and Nelson, 1986), O-scale adiposity rating (Ross et al., 1987) and somatotype (Carter, 1980). The performance variables included: • 7 strength measurements, i.e. 5 cable tensiometer tests (Clarke and Munroe, 1970) and 2 grip strengths; • 17 flexibility measures, i.e. 9 flexometer tests (Leighton, 1983), the sit-reach test (Scott and French, 1959) and a set of 7 field tests for gymnasts developed by Bajin (1978); • 3 endurance measures (bar dips, chin-ups, leg lifts); • 2 power measurements, i.e. the standing broad jump and the vertical jump (Mathews and Fox, 1976); • 2 speed measurements, i.e. a 20-m run and simple total body reaction time; and • 1 balance test on the stabilometer.
Previous injuries, i.e. the total number of injuries sustained prior to the surveillance period as reported by the gymnasts in a background information questionnaire, was also included in the comparisons between the drop-outs and the continuing gymnasts. Data analysis The means and standard deviations for the two categories of gymnasts were calculated and the differences tested through t-tests. The number of variables was then reduced (a) by eliminating the base variables for any kept composite variables (e.g. the six skinfold measurements) or deleting the composite or derived variables for any kept base variables (e.g. ectomorphic somatotype component), (b) by selecting the variables with the highest eigenvector values from clear groupings of variables by separate principal component analyses of the anthropometric and strength and flexibility variables (e.g. thigh circumference representing the four extremity circumferences), and (c) by selecting one of the groups of variables measuring approximately the same ability (e.g. standing broad jump for power). Further multivariate reduction was then accomplished through principal component analyses (SAS Princomp Procedure) of the 22 selected anthropometric and performance variables. The resulting components are listed in Table 2. Component 1 was a large general factor representing body size, strength, power and speed. The second component was a combination of endurance and flexibility. The body fat component was identifiable from the skinfolds and the ponderal index variables, but also contained a flexibility element. The flexometer and sit-reach tests made up the bulk of the fourth component, and because the gymnastic-specific items were not included, this component was labelled 'extent flexibility'. The fifth principal component was a combination of specific flexibility, balance, speed and arm endurance, while component 6 combined a gymnastic-specific flexibility item with strength, power and balance. The seventh and eighth principal components had speed and balance as main elements respectively. These principal components were then used as variables in discriminant analyses along with the previous injury variable. Theses analyses (SAS Discrim Procedure) served to answer the following question: How distinguishable are continuing gymnasts from those who would drop out during the study on the basis of the principal components and the injury variable? Subsequent stepwise discriminant analyses (SAS Stepdisc Procedure) indicated which of the
Withdrawal predictors in gymnastics
263
Table 1. The variables in the test battery and in the principal component analysis
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
Age Height Weight Sit height Ponderal index" Quetelet index Lengths Leg length Left total arm" Right total arm" Left upper arm Left lower arm Left hand Left total leg" Right total leg Left lower leg Right lower leg Left upper leg Right upper leg Left lower total Right lower total Left foot Widths Biacromial" Bideltoid Bicristal" Bitrochanteric Chest Biepicondylar fern. Ankle Biepicondylar hum. Wrist Hand
Circumferences Ankle Calf Forearm Thigh" Hip Lower abdomen Middle abdomen Upper abdomen" Upper arm, flexed Upper arm, extended Chest Knee Wrist Neck Skinfolds Subscapular Triceps Abdominal Suprailiac Thigh Calf Sum 6 skinfolds" % Body fat O-scale adiposity Cable tensiometer Shoulder flexion Shoulder adduction" Knee extension" Knee flexion Ankle plantar flexion Previous injuries Vital capacity
Somatotype Flexometer Shoulder flexion Shoulder abduction Elbow flex.-ext. Radius-ulnar flex.-ext." Knee flex.-ext. Ankle flexion Trunk flexion Trunk extension" Wrist flex.-ext. Splits Forward right" Forward left Sideways Leg raises Forward left Forward right" Sideways left Sideways right Motor performance Grip strength, left Grip strength, right Sit-reach" 20-m run" Leg lifts" Standing broad jump" Bar dips" Chin-ups" Vertical jump Simple reaction time" Balance"
"These variables were selected for further multivariate analyses. components were most important in discriminating between the future drop-outs and the continuing gymnasts.
Results Anthropometry, body composition and somatotype
The future drop-outs as a group were significantly older, taller and heavier than the continuing gymnasts at the time of the testing (Table 3). This difference remained when the gymnasts were categorized according to competitive level, but largely disappeared within the
264
Lindner et al.
Table 2. Principal components for the set of 22 anthropometric and performance variables and their eigenvectors" Component'
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
%var.' Ponderal index Arm length Leg length Biacromial width Bicristal width Thigh circumf. Upper abdomen circumf. Skinfolds Shoulder adduct. Knee extension Radius-ulnar Trunk extension Spilts, forward right Leg raises, forward right Sit-reach 20-mrun Leg lifts
Standing broad jump Bar dips Chin-ups Simple reaction time Balance
Cl
C2
C3
C4
C5
C6
C7
C8
31.05
16.66
8.67
7.22
5.59
4.74
4.63
3.47
-0.59
0.24 0.23
0.32 0.33 0.27 0.32 0.33
-0.31 -0.22
0.26 0.25
0.28
0.30
-0.20 -0.27
0.30 0.29 flex.
0.51 -0.28
0.30
0.29 0.28
0.41 0.36
0.31 0.39 0.27
0.23 0.52
-0.29
-0.23
0.47
-0.27
0.42
-0.25
0.36 -0.40
0.22
0.42
0.32 0.35
-0.31
0.30
-0.25
0.43
0.30 -0.36 -0.31 -0.22 0.24
-0.38 0.85 0.55
"Only variables with loadings of over 0.20 are listed. *% var., percentage of total variation accounted for. C C1, body size, strength, power and speed; C2, endurance and flexibility; C3, body fat and flexibility; C4, extent flexibility; C5, balance, speed and arm flexibility and strength; C6, gymnastic-specific flexibility and power vs endurance; C7, speed; C8, balance. age intervals. The drop-outs also had broader shoulders in all groups, significantly so in the whole sample and the two competitive levels. There were no significant differences in body fat measurements for any of the groups and no clear trend was apparent. The drop-outs tended to score lower on the mesomorphy somatotype component, with the exception of the two younger age groups, but only for age group 13 was the comparison statistically significant. The mean somatotype for the whole sample was 1.95 (endo), 4.18 (meso) and 3.29 (ecto), while the comparison between the dropouts and the continuing gymnasts yielded somatotypes of 1.89-4.03-3.44 and 2.04-4.29-3.22 respectively. The means of the previous injuries variable tended to be higher (but not significantly so) for the drop-outs in all comparisons except in the age 9 interval.
Table 3. Means (x) and standard deviations (S.D.) of selected anthroporaetric, somatotype and body composition variables for drop-outs (D) and continuing gymnasts (C) All
Downloaded by [University of Toronto Libraries] at 11:20 30 January 2015
Variable
Low-level
High-level
D (" = 27)
C (n=41)
D (n = 13)
C (n = 30)
D (n=14)
C (n = l l )
12.6* 1.7
10.6 1.4
12.1** 1.5
10.2 1.2
13.1 1.7
11.9 1.3
147.5* 10.0
136.5 9.6
146.9** 11.4
135.2 9.3
148.1ft 7.7
39.0* 6.9
31.7 6.7
37.4** 6.6
30.8 6.6
292.2* 21.9
259.6 26.1
291.9* 21.9
451.1* 34.6
416.4 38.9
48.4 9.2
Age 9 group D (n = 4)
C ("=23:
Age 11 group D )
Age 13 group
C ( i = i 4 ;1
D (n = 16)
C (n = 4)
Age (years) X
S.D.
0.7
9.6 0.8
11.2 0.6
11.6 0.5
13.7 1.1
13.2 0.7
139.8 10.2
134.0 10.9
130.6 6.3
142.9 6.3
142.7 7.8
152.9 5.2
148.4 6.7
40.6ft 7.0
34.0 6.6
30.3 5.4
27.9 3.9
34.9 6.1
34.8 4.7
42.6 4.8
42.5 8.7
253.8 26.7
292.5ft 22.7
275.4 17.0
263.5 16.8
250.4 22.6
284.0 24.3
264.6 24.5
302.9 12.8
295.5 14.9
437.9 31.1
412.5 40.7
463.3f 34.2
427.0 33.1
423.8 37.4
396.2 28.5
437.7 38.0
434.6 25.7
463.8 27.7
469.3 55.0
50.0 15.7
48.9 10.6
52.7 16.7
47.9 8.1
42.4 9.5
50.3 16.2
48.8 14.5
49.7 8.3
52.1 16.9
47.3 8.0
49.4 22.1
11.8 2.4
11.9 3.3
12.4 3.1
12.4 3.4
11.3 1.5
10.4 2.2
13.3 5.2
11.7 3.1
11.7 1.8
12.3 3.5
11.5 1.7
11.0 3.8
1.9 0.6
2.0 0.9
1.9 0.7
2.2 1.0
1.9 0.5
1.7 0.5
2.1 1.3
2.0 0.8
1.9 0.3
2.1 1.1
1.9 0.5
2.0 0.7
4.0 0.8
4.3 0.8
3.9 0.9
4.2 0.7
4.1 0.6
4.6 0.6
4.5 1.1
4.4 0.6
4.3 0.8
4.0 0.7
0.7
3.8ft
4.8 1.0
3.4 0.9
3.2 0.9
3.7 1.0
3.2 1.0
3.2 0.8
3.2 0.5
3.0 1.1
3.2 0.8
3.2 1.2
3.5 1.0
3.7 0.8
2.8 1.3
1.44 1.12
1.00 1.16
1.08 0.76
0.73 0.99
1.79 1.31
1.73 1.35
0.50 0.52
0.67 0.93
1.57 1.27
1.36 1.22
1.63 1.09
1.50 1.92
10.4ft
Height (cm) X
S.D.
.
1o* 1
Weight (kg) X
S.D.
8
Biac. width (mm)" X
S.D.
Thigh circumf. (mm) X
S.D.
Skinfolds* X
S.D.
% Fat X
S.D.
Endomorph.
c
X
S.D.
Mesomorph.11 X
S.D.
Ectomorph.6 X
S.D.
Prev. inj. (no./ X
S.D.
"Biac. width, biacromial width; 'skinfolds, sum of six skinfolds; "endomorph., endomorphic somatotype component; ''mesomorph., mesomorphic somatotype component; "ectomorph., ectomorphic somatotype component; •'prev. inj., previous injuries. *P
