This article was downloaded by: [University of Wyoming Libraries] On: 26 August 2013, At: 15:11 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
The relationships between technique variability and performance in discus throwing a
b
c
a
Boyi Dai , Steve Leigh , Hanjun Li , Vicki S. Mercer & Bing Yu
a
a
University of North Carolina at Chapel Hill, Division of Physical Therapy, Chapel Hill, United States b
Montclair State University, Department of Exercise Science and Physical Education, Montclair, United States c
Beijing Sport University, College of Exercise Science, Beijing, China Published online: 11 Oct 2012.
To cite this article: Boyi Dai , Steve Leigh , Hanjun Li , Vicki S. Mercer & Bing Yu (2013) The relationships between technique variability and performance in discus throwing, Journal of Sports Sciences, 31:2, 219-228, DOI: 10.1080/02640414.2012.729078 To link to this article: http://dx.doi.org/10.1080/02640414.2012.729078
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Sports Sciences, 2013 Vol. 31, No. 2, 219–228, http://dx.doi.org/10.1080/02640414.2012.729078
The relationships between technique variability and performance in discus throwing
BOYI DAI1*, STEVE LEIGH2, HANJUN LI3, VICKI S. MERCER1, & BING YU1 University of North Carolina at Chapel Hill, Division of Physical Therapy, Chapel Hill, United States, 2Montclair State University, Department of Exercise Science and Physical Education, Montclair, United States, and 3Beijing Sport University, College of Exercise Science, Beijing, China
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
1
(Accepted 7 September 2012)
Abstract Discus throwing performance is associated with technique variables. However, the relationships between technique variability and performance in discus throwing are unknown. Video clips of three throws each by 18 male and 15 female discus throwers were collected during a competition. Two-dimensional coordinate data were manually digitised and transferred into three-dimensional coordinates of body landmarks. Eight joint angles were calculated and their values were time-normalised to 101 frames from the initiation to the end of a throw. The mean standard deviations of 101 frames for eight joint angles among the three throws were calculated as the variability measurement. The best official distance of three throws was used as the performance measurement. For male throwers, the canonical correlations between technique variability and performance were significant (P 5 0.1). Pearson correlations showed that the performance was significantly negatively correlated with the mean standard deviations of arm-shoulder separation angle (r ¼ 70.57, P ¼ 0.02), hipshoulder separation angle (r ¼ 70.63, P 5 0.01), left hip flexion angle (r ¼ 70.5, P ¼ 0.03), and trunk forward-backward tilt angle (r ¼ 70.57, P ¼ 0.02). For female throwers, the canonical correlations between technique variability and performance were not significant (P 4 0.1). Male discus throwers with a longer throwing distance had smaller variability in reproducing techniques. Reducing motion variability may be an important goal for discus training in males.
Keywords: variability, performance, discus throwing, biomechanics, kinematics
Introduction Discus throwing is a physically and technically demanding event in track and field that requires athletes to perform a sequence of high speed movements in a limited space. The performance of a discus throw is measured by official distance, which is the sum of vacuum flight distance, aerodynamic distance, and distance lost at release (Bartlett, 1992; Hay & Yu, 1995). Previous investigators demonstrated that release speed was the primary determinant of discus throwing performance (Bartlett, 1992; Hay & Yu, 1995; Leigh, Gross, Li, & Yu, 2008; Leigh & Yu, 2007). Discus throwing performance was also associated with kinematic measures such as hip-shoulder separations, shoulder-arm separations, and trunk tilt angles (Leigh et al., 2008; Leigh & Yu, 2007); and kinetic measures such as ground reaction forces and lower extremity joint moments (Yu,
Broker, & Silvester, 2002). Previous investigators further demonstrated intra-individual variations in kinematic and kinetic measures of discus throwing techniques, which indicate variability in motor control in discus throwing (Bauer & Schollhorn, 1997; Leigh, Liu, Hubbard, & Yu, 2010). Variability associated with a motor task includes both outcome variability and movement variability. Outcome variability typically decreases as a function of practice and incremental improvement in skill (Newell & Corcos, 1993). However, the relationships between movement variability and skill level may be largely dependent on the task (Broderick & Newell, 1999; Button, MacLeod, Sanders, & Coleman, 2003; Wilson, Simpson, van Emmerik, & Hamill, 2008). Motor programming and dynamic systems are two major motor control theories. Motor programming theory suggests that once a motor task has been
Correspondence: Bing Yu, University of North Carolina at Chapel Hill, Division of Physical Therapy, CB# 7135 Bondurant Hall, Chapel Hill, 27599-7135 United States. E-mail:
[email protected] *The author is currently at: University of Wyoming, Division of Kinesiology and Health, Laramie, United States. Ó 2013 Taylor & Francis
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
220
B. Dai et al.
learned, the generalised motor programme, including sequencing, relative timing, and relative force, will be relatively constant, while parameters such as overall duration and overall force can be varied according to changes in task demands (Schmidt, 1975; Schmidt & Lee, 2005). According to motor programming theory, decreased variability is associated with increased efficiency when performing a motor task (Schmidt & Lee, 2005). Dynamic systems theory, on the other hand, assumes that the movement system is flexible and changes over time between stable and unstable patterns (Kelso & Scho¨nera, 1988). A consistent outcome can be achieved by different combinations of degrees of freedom (Davids, Glazier, Araujo, & Bartlett, 2003). A certain amount of variability gives the system flexibility in response to changes in movement constraints (Newell & Vaillancourt, 2001). In general, motor programming is more applicable to discrete and closed skills, while dynamic systems theory is applicable to continuous and open skills (Schmidt, 2003). Discrete movements (e.g., kicking) have a recognisable beginning and end, while continuous movements (e.g., running) have no recognisable beginning or end (Schmidt & Lee, 2005). Open skills (e.g., playing a basketball game) are those for which the environment is constantly changing, while closed skills (e.g., bowling) are those for which the environment is relatively stable (Schmidt & Lee, 2005). During discus throwing, a throw is initiated with maximum back swing and ended with the release of the discus. Environmental factors, such as wind speed and direction, can influence the thrower’s choice of techniques. However, the techniques are most likely programmed before the throw and minimally influenced by the environment during the throw. In addition, the environment is relatively constant during the same competition. Therefore, discus throwing may be considered a discrete and closed task and therefore may be best understood within the context of motor programming theory. Although variability is viewed as an important aspect of motor control, the effects of technique variability on sports performance have not received adequate research attention, particularly for discus throwing. Previous researchers have examined relationships between technique variability and performance in golf (Bradshaw et al., 2009), basketball (Broderick & Newell, 1999; Button et al., 2003), javelin throwing (Bartlett, Mu¨ller, Lindinger, Brunner, & Morriss, 1996; Bartlett, Wheat, & Robins, 2007), and triple jump (Wilson et al., 2008). Considering the presence of intra-individual variation in discus throwing techniques and the uncertain association between variation in technique and performance of a given sport, the purpose of this
study was to investigate the relationships between technique variability and performance of discus throwing. It was hypothesised that throwing performance would be negatively correlated with variability of technique measures, in accordance with motor programming theory.
Methods Participants This study was approved by the University of North Carolina at Chapel Hill Institutional Review Board. Eighteen male and 15 female discus throwers competing in the 2010 USA Outdoor Track & Field Championships were included in this study. All participants were right-handed throwers. All three trials of each thrower in the preliminary round were included in analysis of this study. The reason to include the preliminary round was to maximise the number of participants, because only eight throwers were eligible for the second round. In addition, for both male and female throwers, the eight finalists had similar performance between their first round and second round. Data collection Data collection and data reduction procedures have been described in detail in previous studies (Leigh et al., 2008, 2010; Leigh & Yu, 2007). Two high definition video camcorders were used to record throwers’ performances at a sampling frequency of 60 frames s71. The angle between the optical axes of the two camcorders was approximately 90 degrees (Figure 1). The two cameras were calibrated using a calibration frame with 24 calibration points for a space 2.5 m long, 2 m wide, and 2.5 m high that covered the space in which all throwers performed their throws. Five markers were placed in the throwing circle to establish a fixed world reference frame (Figure 1). Data reduction Twenty-one body landmarks and the centre of the discus were manually digitised from maximum backswing to two frames after discus release using a Motus videographic data acquisition system (Peak Performance Technology, Inc., Englewood, CO). The calibration points and fixed world reference frame markers captured in a calibration video frame were also digitised. The digitised two-dimensional coordinates during discus throws were synchronised using five critical events including right foot off, left foot off, right foot down, left foot down, and release of the discus (Hay & Yu, 1995). The direct linear
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
Variability and performance in discus throwing
221
Figure 1. Positions of camcorders and fixed world reference frame.
transformation (Abdel-Aziz & Karara, 1971) procedure was used to transfer the two-dimensional coordinates to real-time three-dimensional coordinates at a frequency of 60 frames s71. The threedimensional coordinates were then filtered through a Butterworth filter at a low-pass frequency of 7.14 Hz (Yu & Andrews, 1998). The data synchronisation, transformation, and filter were performed using a custom software (MotionSoft, Chapel Hill, NC). The initiation of a throw was defined as the first frame the discus’s speed consistently reached more than 1 m s71 after maximum backswing. The end of a throw was visually identified as the frame in which the discus left the thrower’s hand. The average frames from maximum back swing to discus attaining sustained speed greater than 1 m s71 are approximately seven frames for both male and female throwers. By using this method to quantify the initiation and the end of a throw, the variance in the total number of frames across three throws for each thrower was usually one or two frames. In view of the key components of techniques and the major joints contributing to discus delivery, the variables of arm-shoulder separation angle, arm-shoulder elevation angle, shoulder-hip separation, right hip flexion angle, left hip flexion angle, right knee flexion angle, left knee flexion angle, and trunk forward-backward tilt angle were selected for study and were calculated from the initiation to the end of a throw at each realtime output frame (Leigh et al., 2008). For joint angle calculations, local reference frames of upper trunk, lower trunk, right thigh, left thigh
and a moving global reference frame were established. For upper trunk, the z-axis was defined as the unit-vector passing from the mid-point between left and right hips to the mid-point between left and right shoulders. An intermediate axis was defined as the unit-vector passing from right shoulder to left shoulder. The x-axis was defined as the cross product of intermediate axis and z-axis. The y-axis was defined as the cross product of z-axis and x-axis. For lower trunk, the definitions of axes were the same as for upper trunk, except that the intermediate axis was defined as the unit-vector passing from right hip to left hip. For left and right thighs, the z-axis was defined as the unit-vector passing from knee to hip. An intermediate axis was defined as the unit-vector passing from ankle to hip. The y-axis was defined as the cross product of z-axis and intermediate axis. The x-axis was defined as the cross product of y-axis and z-axis. For the moving global reference frame, the origin was defined as the mid-point of left and right hips. The z-axis was defined as the unit-vector of the fixed world axis perpendicular to the ground with upward positive. An intermediate axis was defined as the unit-vector passing from right hip to left hip. The x-axis was defined as the cross product of intermediate axis and z-axis. The y-axis was defined as the cross product of z-axis and x-axis. An arm vector was defined as the vector passing from right shoulder to right wrist. The arm-shoulder separation angle (Figure 2) was calculated as the angle between the projection of the arm vector on the x-y plane of the upper trunk reference frame and
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
222
B. Dai et al.
Figure 2. Definitions of arm-shoulder separation angle and shoulder-hip separation angle.
the y-axis of the upper trunk reference frame. The arm-shoulder elevation angle (Figure 3) was calculated as the angle between the projection of the arm vector on the y-z plane of upper trunk reference frame and the y-axis of the upper trunk reference frame. A hip vector was defined as the vector passing from right hip to left hip. The shoulder-hip separation angle (Figure 2) was calculated as the angle between the projection of the hip vector on the x-y plane of the upper trunk reference frame and the y-axis of the upper trunk reference frame. Cardan angles between the lower trunk and left/ right thigh reference frames were calculated with a rotation order of extension-flexion (y-axis), adduction-abduction (x-axis), and internal rotation-external rotation (z-axis). The extension-flexion angle was referred to as hip flexion angle (Figure 4) and used for further analysis. A shank vector was defined as the vector passing from the ankle to the knee. A thigh vector was defined as the vector passing from the knee to the hip. The right and left knee flexion angle (Figure 5) were calculated as two-dimensional angles between the shank and thigh vectors. Cardan angles between lower trunk and moving global reference frames were calculated with a rotation order of extension-flexion (y-axis), adduction-abduction (x-axis), and internal rotationexternal rotation (z-axis). The extension-flexion angle was referred to as trunk forward-backward tilt angle (Figure 6) and used for further analysis. The joint angles were time-normalised to 101 points from the initiation to the end of a throw using a linear interpolation method. For each joint angle, the standard deviations (s) among three trials at each normalised time point were calculated. The mean of the standard deviations of 101 normalised time points was calculated for each angle as a
Figure 3. Definitions of arm-shoulder elevation angle.
Figure 4. Definitions of right and left hip flexion angle.
measurement of variability (Li, Haddad, & Hamill, 2005). The best official distance of three throws for each thrower was used as the measurement for performance, because this determined the thrower’s final place in the competition. To quantify the variability resulting from manual digitising errors, one throwing trial from among all males and one throwing trial from among all females were randomly selected and digitised twice. The average digitising variability for each of these two trials is presented in Table I. The calculations of joint angles and variability measurements were conducted in MATLAB (MathWorks Inc., PA, USA).
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
Variability and performance in discus throwing
Figure 5. Definitions of right and left knee flexion angle.
223
variability measurements (Tabachnick & Fidell, 2007). In a canonical correlation analysis, canonical variates are constructed for a set of independent variables and a set of dependent variables to maximise the correlations between independent canonical variates and dependent variates with a goal to decrease the data dimensions among variables (Tabachnick & Fidell, 2007). For the current study, a one-dimensional canonical variate was constructed for variability measurements to maximally explain the variance in performance. A finding of a significant canonical correlation was interpreted as suggesting significant correlations between performance and variability measurements; analysis then proceeded with Pearson correlation analysis between performance and individual variability measurements to explore the discrete correlations. If a non-significant canonical correlation was found, this suggested non-significant correlations between performance and variability. Because the canonical correlation analyses indicated the correlation between performance and overall variability, significant correlation between performance and individual variability measurement might be found when the canonical correlation analyses were approaching significance. Therefore, a priori Type I error rate was set at 0.1 for canonical correlation analyses. A priori Type I error rate was set at 0.05 for Pearson correlation analysis. Statistical analyses were conducted in SPSS 16.0 (SPSS, IL, USA). Results
Figure 6. Definitions of trunk forward-backward tilt angle.
Statistical analysis To reduce the influence of gender, strength, and discus weights (Leigh & Yu, 2007), males and females were analysed separately in the current study. To reduce the number of statistical tests, canonical correlation analyses were conducted as a screening tool to determine the relationship between performance and the linear combinations of
The variability across three throwing trials was approximately three times the value of digitising variability for all joint angles (Tables I, II). For male throwers, the canonical correlation between performance and variability measurements was significant (P ¼ 0.067). The correlations between the canonical variate and individual variability measurements were all negative in males (Table III). Pearson correlations showed that the performance was significantly negatively correlated with the mean standard deviations of arm-shoulder elevation angle (r ¼ 70.58, P ¼ 0.01, Figure 7, Table IV), shoulder-hip separation angle (r ¼ 70.63, P 5 0.01, Figure 8), left hip flexion angle (r ¼ 70.5, P ¼ 0.03, Figure 9), and trunk forward-backward tilt angle (r ¼ 70.57, P ¼ 0.01, Figure 10) in males. For female throwers, the canonical correlation between performance and variability measurements was not significant (P ¼ 0.21). The correlations between canonical variate and individual variability measurement were negative except for right knee flexion angle in females (Table III). The Pearson correlation coefficients between performance and individual variability
B. Dai et al.
224
Table I. Joint angle variability across two identical trials (digitising variability). ASS (Deg)
ASE (Deg)
SHS (Deg)
RHF (Deg)
LHF (Deg)
RKF (Deg)
LKF (Deg)
TFBT(Deg)
1.0
1.1
2.5
2.0
1.6
1.5
1.0
0.9
Note: ASS: arm-shoulder separation angle; ASE: arm-shoulder elevation angle; SHS: shoulder-hip separation; RHF: right hip flexion angle; LHF: left hip flexion angle; RKF: right knee flexion angle; LKF: left knee flexion angle; TFBT: trunk forward-backward tilt angle.
Table II. Mean + standard deviations of best official distance and joint angle variability across three throws.
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
Males Females
Distance (m)
ASS (Deg)
ASE (Deg)
SHS (Deg)
RHF (Deg)
LHF (Deg)
RKF (Deg)
LKF (Deg)
TFBT (Deg)
56.7 + 2.7 54.1 + 5.7
4.5 + 0.9 5.2 + 1.0
3.0 + 0.8 4.1 + 1.2
7.0 + 1.0 6.9 + 1.2
5.3 + 0.8 5.9 + 1.3
4.7 + 0.5 4.2 + 0.8
4.5 + 0.9 4.2 + 0.6
4.1 + 0.7 4.1 + 0.9
2.7 + 0.5 2.4 + 0.6
Note: ASS: arm-shoulder separation angle; ASE: arm-shoulder elevation angle; SHS: shoulder-hip separation; RHF: right hip flexion angle; LHF: left hip flexion angle; RKF: right knee flexion angle; LKF: left knee flexion angle; TFBT: trunk forward-backward tilt angle.
Table III. Pearson correlations coefficients between the canonical variate for variability and individual joint angle variability.
Males Females
ASS
ASE
SHS
RHF
LHF
RKF
LKF
TFBT
70.26 70.24
70.68 70.64
70.74 70.17
70.53 70.13
70.59 70.07
70.46 0.09
70.28 70.22
70.67 70.43
Note: ASS: arm-shoulder separation angle; ASE: arm-shoulder elevation angle; SHS: shoulder-hip separation; RHF: right hip flexion angle; LHF: left hip flexion angle; RKF: right knee flexion angle; LKF: left knee flexion angle; TFBT: trunk forward-backward tilt angle.
Figure 7. Arm-shoulder elevation angle variability as a function of best throwing distance in male throwers.
measurement were negative except for right knee flexion angle in females (Table V). Performance was only significantly negatively correlated with the mean standard deviations of arm-shoulder elevation angle (r ¼ 70.54, P ¼ 0.04).
Discussion The hypothesis of the current study was supported by the study results for males, but not for females. For males, canonical correlation analysis suggested that performance was negatively correlated with the
Variability and performance in discus throwing
225
Table IV. Pearson correlation coefficients and P values between best official distance and individual joint angle variability in males.
Pearson correlation coefficients P values
ASS
ASE
SHS
RHF
LHF
RKF
LKF
TFBT
70.22 0.38
70.58 0.01
70.63 50.01
70.45 0.06
70.5 0.03
70.39 0.11
70.24 0.34
70.57 0.01
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
Note: ASS: arm-shoulder separation angle; ASE: arm-shoulder elevation angle; SHS: shoulder-hip separation; RHF: right hip flexion angle; LHF: left hip flexion angle; RKF: right knee flexion angle; LKF: left knee flexion angle; TFBT: trunk forward-backward tilt angle.
Figure 8. Shoulder-hip separation angle variability as a function of best throwing distance in male throwers.
Figure 9. Left hip flexion angle variability as a function of best throwing distance in male throwers.
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
226
B. Dai et al.
Figure 10. Trunk forward-backward tilt angle variability as a function of best throwing distance in male throwers.
Table V. Pearson correlation coefficients and P values between best official distance and individual joint angle variability in females.
Pearson correlation coefficients P values
ASS
ASE
SHS
RHF
LHF
RKF
LKF
TFBT
70.21 0.46
70.54 0.04
70.14 0.61
70.11 0.69
70.06 0.83
0.08 0.78
70.19 0.51
70.37 0.18
Note: ASS: arm-shoulder separation angle; ASE: arm-shoulder elevation angle; SHS: shoulder-hip separation; RHF: right hip flexion angle; LHF: left hip flexion angle; RKF: right knee flexion angle; LKF: left knee flexion angle; TFBT: trunk forward-backward tilt angle.
combinations of the variability of shoulder, hip, knee, and trunk angles. The Pearson correlation coefficients between performance and individual variability measurements were all negative, with armshoulder elevation angle, shoulder-hip separation angle, left hip flexion angle, and trunk forwardbackward tilt angle variability being statistically significant. The negative correlations indicate that male throwers with longer official distances had lower joint angle variability across three throwing trials. For females, although most canonical correlation loadings were negative, a lack of statistical significance suggested that the correlation was not strong in this participant group. Previous investigators have demonstrated that during certain sports manoeuvres, skilled individuals had lower intra-individual variability in key components of techniques compared to less skilled individuals, and that variability decreased with practice and learning (Bradshaw et al., 2009; Button et al., 2003). Bradshaw et al. (2009) showed that skilled golfers had lower standard deviations in club-head velocity and key joint kinematics than unskilled golfers during a golf swing. Button et al. (2003)
found increases in skill levels were associated with decreased variability in wrist and elbow coordination in golfers. Broderick and Newell (1999) found decreased variability in linear and angular displacements of the upper extremity with increased skill level in basketball bouncing. These findings support motor programming theory, which focuses on invariant movement characteristics for each class of motor skills and considers variability as noise in the motor output. However, some other investigators suggested elite athletes might increase variability to give the body flexibility in response to changes in movement constraints (Bartlett et al., 1996; Wilson et al., 2008). Bartlett et al. (1996, 2007) reported that joint variability was greater for novice and elite javelin throwers than for club javelin throwers whose skill level was between novice and elite throwers. A ‘‘U’’ shaped relationship between joint coordination variability and skill level was also observed for triple jumpers (Wilson et al., 2008). The two most skilled and the one least skilled triple jumpers demonstrated higher joint coordination variability than the two middle skilled triple jumpers (Wilson et al., 2008). These findings support the dynamic systems theory,
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
Variability and performance in discus throwing which considers variability as the component that triggers changes in the way movements are organised and gives the movement system flexibility in response to external perturbations. In the current study, the results for males were in agreement with the motor programming theory and support the notion of better movement consistency in skilled athletes. Discus throwing is a discrete, closed skill, and the environment has relatively little influence on the throwers’ performance. A motor programme for discus throwing may be formulated in elite male throwers after long-term practice and training. Consistent with this idea, the intra-individual technique variability was generally small, suggesting high consistency in elite male throwers. Therefore, discus throwing is a sport for which motor programming theory may be directly applicable. Athletes with better performance should have better learning of the motor programme and less noise coming from the internal system during movement execution. An important point to keep in mind, however, is that these data were collected during actual competition, in which the goal was to achieve maximum performance, presumably by replicating optimal trained techniques. If the goal had been to learn new techniques, the throwers would have been likely to increase variability and release movement degrees of freedom. Javelin throwing and triple jump are also closed tasks, but a ‘‘U’’ shape relationship between joint coordination variability and skill level was observed in previous studies of these sports (Bartlett et al., 1996; Wilson et al., 2008). In the current study, the relationship between performance and variability in male throwers is more like a ‘‘\’’ shape instead of ‘‘U’’ shape. The discrepancies among studies may result from the methods used to quantify movement variability. In the current study, the movement variability was evaluated using mean standard deviations, while some previous investigators used phase angle and vector coding to quantify variability (Button et al., 2003; Wilson et al., 2008). Phase angle and vector coding calculated the movement variability between two coupling segments or joints as compared to calculating the mean standard deviation for a single joint (Button et al., 2003; Wilson et al., 2008). Therefore, the variability in the current study refers to single joint angle variability rather than joint coordination variability. In addition, previous investigators generally compared athletes with high and low skills, while the athletes in the current study were all elite athletes. The relationship between technique variability and performance in low and high skilled athletes might not hold in elite athletes with relatively small performance differences. Furthermore, although previously mentioned tasks were all closed tasks, tasks such as basketball free throws are
227
performed for accuracy while tasks such as javelin throwing and discus throwing are performed for distance. These differences in sports characteristics, athletes’ competition levels, data collection contexts, and methods used to quantify variability could all contribute to the differences among studies. The finding of no significant correlation between performance and variability in females suggests that variability may be not only sports-specific but also gender-specific. Previous investigators found that the magnitude of arm-shoulder separation angle, shoulder-hip separation, and trunk forward-backward tilt angles at critical instants were associated with performance in females, while only shoulderhip separation angle was correlated with performance in males (Leigh et al., 2008; Leigh & Yu, 2007). The authors suggested that females tend to rely on effective technique, while males may have higher inter-participant consistency in technique and greater dependence on physical strength to achieve better performance. The gender disparity may be caused by different throwing styles, coaching theories, and physical characteristics between males and females. Further analysis in the current study also revealed larger inter-participant variability in joint angles in females compared to males. If male throwers all tend to use similar techniques, then perhaps there is less opportunity for intraparticipant variability in male throwers, and they may exhibit greater consistency in replicating desired techniques as their skills improve. If this is the case, the likelihood of finding negative correlations between variability and performance would be higher in males as compared to females. However, it should be noted that although the statistical test was not significant, the canonical loadings were negative for most variability measurements for females. An increase in sample size might detect a significant correlation between variability and performance in females. The current study had several limitations that may need to be noticed. This was a cross-sectional correlation study, so no cause-effect relationship can be inferred. It is still unknown if decrease or increase in variability can induce changes in performance. Future studies with longitudinal design might overcome this limitation. Although we found that the variability in those technical parameters due to digitising errors was small, and that the majority of intra-participant variability in the current study came from the technique variability, we did not have any information regarding the effects of digitising errors on the correlations observed in this study. In addition, some previous investigators used other methods such as vector coding to quantify coordination variability between two joints (Button et al., 2003; Wilson et al., 2008). Because the major
228
B. Dai et al.
Downloaded by [University of Wyoming Libraries] at 15:11 26 August 2013
application of the current study was to provide insight into discus training, we used mean standard deviations of a single joint angle instead of variability in coordination between joints, with an aim to provide straightforward information for athletes and coaches. This study was only a start of the research on this topic. Future studies need to study the correlations of the variability in those technical parameters at specific instants. Future studies with other data reduction techniques can provide a more comprehensive understanding of relationships between variability and performance in discus throwing. Conclusion Negative correlations between technique variability and performance were found in elite male, but not female, discus throwers during competition. The findings for the male throwers supported motor programming theory, which suggested male athletes with better performance had better consistency in reproducing trained techniques. The gender disparity might be caused by different throwing styles, coaching theories, and physical characteristics between males and females. Reducing motion variability may be an important goal for discus training in males.
References Abdel-Aziz, Y.I., & Karara, H.M. (1971). Direct linear transformation from comparator coordinates into object space coordinates in close-range photogrammetry. In Proceedings of the Symposium on Close-Range Photogrammetry (pp. 1–18). Falls Church, VA: American Society of Photogrammetry. Bartlett, R.M. (1992). The biomechanics of the discus throw: A review. Journal of Sports Sciences, 10, 467–510. Bartlett, R.M., Mu¨ller, E., Lindinger, S., Brunner, F., & Morriss, C. (1996). Three-dimensional evaluation of the kinematic release parameters for javelin throwers of different skill levels. Journal of Applied Biomechanics, 12, 58–71. Bartlett, R.M., Wheat, J., & Robins, M. (2007). Is movement variability important for sports biomechanists? Sports Biomechanics, 6, 224–243. Bauer, H.U., & Schollhorn, W. (1997). Self-organizing maps for the analysis of complex movement patterns. Neural Processing Letters, 5, 193–199.
Bradshaw, E.J., Keogh, J.W., Hume, P.A., Maulder, P.S., Nortje, J., & Marnewick, M. (2009). The effect of biological movement variability on the performance of the golf swing in high- and low-handicapped players. Research Quarterly for Exercise and Sport, 80, 185–196. Broderick, M.P., & Newell, K.M. (1999). Coordination patterns in ball bouncing as a function of skill. Journal of Motor Behavior, 31, 165–188. Button, C., MacLeod, M., Sanders, R., & Coleman, S. (2003). Examining movement variability in the basketball free-throw action at different skill levels. Research Quarterly for Exercise and Sport, 74, 257–269. Davids, K., Glazier, P., Araujo, D., & Bartlett, R. (2003). Movement systems as dynamical systems: The functional role of variability and its implications for sports medicine. Sports Medicine, 33, 245–260. Hay, J.G., & Yu, B. (1995). Critical characteristics of technique in throwing the discus. Journal of Sports Sciences, 13, 125–140. Kelso, J.A.S., & Scho¨nera, G. (1988). Self-organization of coordinative movement patterns. Human Movement Science, 7, 27–46. Leigh, S., Gross, M.T., Li, L., & Yu, B. (2008). The relationship between discus throwing performance and combinations of selected technical parameters. Sports Biomechanics, 7, 173–193. Leigh, S., Liu, H., Hubbard, M., & Yu, B. (2010). Individualized optimal release angles in discus throwing. Journal of Biomechanics, 43, 540–545. Leigh, S., & Yu, B. (2007). The associations of selected technical parameters with discus throwing performance: A cross-sectional study. Sports Biomechanics, 6, 269–284. Li, L., Haddad, J.M., & Hamill, J. (2005). Stability and variability may respond differently to changes in walking speed. Human Movement Science, 24, 257–267. Newell, K.M., & Corcos, D.M. (1993). Variability and motor control. Chicago, IL: Human Kinetics. Newell, K.M., & Vaillancourt, D.E. (2001). Dimensional change in motor learning. Human Movement Science, 20, 695–715. Schmidt, R.A. (1975). A schema theory of discrete motor skill learning. Psychological Review, 82, 225–260. Schmidt, R.A. (2003). Motor schema theory after 27 years: Reflections and implications for a new theory. Research Quarterly for Exercise and Sport, 74, 366–375. Schmidt, R.A., & Lee, T.D. (2005). Motor control and learning: A behavioral emphasis (4th ed.). Champaign, IL: Human Kinetics. Tabachnick, B.G., & Fidell, L.S. (2007). Using multivariate statistics (5th ed.). Needham Heights, MA: Allyn & Bacon. Wilson, C., Simpson, S.E., van Emmerik, R.E., & Hamill, J. (2008). Coordination variability and skill development in expert triple jumpers. Sports Biomechanics, 7, 2–9. Yu, B., & Andrews, J.G. (1998). The relationship between free limb motions and performance in the triple jump. Journal of Applied Biomechanics, 14, 223–237. Yu, B., Broker, J., & Silvester, L.J. (2002). A kinetic analysis of discus-throwing techniques. Sports Biomechanics, 1(1), 25–46.
