Football (also known as soccer) is one of the most popular team sports worldwide (Katis & Kellis, 2007) with hundreds of millions purported to play (Masuda et al, 2005) and in accordance is watched on 6 continents (Ekstrand, 1994). Due to this popularity, football is a widely researched area with the volume of literature extensive. Various research programmes have been undertaken in the area of football kick biomechanics with a range of parameters being measured and analysed, in an attempt to understand the fundamental skills required by the sport, especially the maximal soccer kick (Lees & Nolan, 1998). Although the field is widely researched gaps still transpire. One of these gaps is gender differences; little research is documented on the kick biomechanics of women’s football as said by Barfield et al (2002), who states ‘the rapid rise in female participation in soccer worldwide has not been followed by a corresponding increase in the number of studies biomechanically that target female kicking patterns to determine if differences exist between males and females’. Lee and Nolan (1998) state that success in football depends on kicking performance, with new aspects of this being identified (Kathis & Kellis, 2007). Shan and Westerhoff (2005) believe that the scientific understanding of the sport is not yet on the same echelon as its practice, subsequently its partakers acquire their skills not through research based instruction but through individual experience; suggesting that biomechanical feedback may facilitate an athlete further.
Biomechanical kicking success in football has been measured predominately by maximum ball velocity (Markovic et al, 2006) with Dorge et al (2002) stating it could be this speed that is particularly important when kicking towards goal. When kicking a ball, players will use the most appropriate form dependant on the intent and nature of the outcome (Numone et al, 2002) and according to a study by Grant et al (1998), who analysed data from the 1998 World Cup, the instep kick (IK) (see figure 1) and sidefoot kick are the most commonly used techniques to score. The ball velocity of the maximum IK is the main indicator in kicking performance (Orloff et al, 2008) and has been said to be the result of various factors including technique (Lees and Nolan, 1998), gender (Barfield et al, 2002), muscle strength and power of players (De Proft et al, 1988; Dutta & Subraminium, 2002).
The IK is a fundamental skill that is used on many occasions during football (see figure 1), with Orloff et al (2008) stating that the mechanics in instep kicking are critical in determining kick performance. Transfer of momentum from the thigh to the leg is believed to play an important role in instep kicking, however these claims have not been conclusively quantified (Dunn & Putnam, 1988).
The IK involves a sequence of momentum from proximal (thigh) to distal (shank and foot) body segments in the kicking limb as it is a swing action (Barfield et al, 2002) that should be a natural fluid motion (Clagg et al, 2009). When a kick is performed the proximal segment initiates the movement taking the kicking leg backwards, with the distal segment lagging behind, forward movement of the leg occurs when the proximal segment has reached its potential at backswing and is brought forward whilst the knee continues to flex (Wickstrom, 1975; Dorge et al, 2002). This is followed by a deceleration of the proximal segment due to motion dependant moments from the shank (Putnam, 1991); upon ball impact the proximal segment is almost stationary, at which point the distal segment is accelerating and vigorously extending about the knee to almost full extension at ball impact (Wickstrom, 1975) (see figure 2). At the point of contact, of instep to ball, powerful kickers keep the foot/ankle complex locked and plantarflexed, as a consequence the forces that propel the ball are maximised (Hay, 1996; Tsaousidid and Zatsiorsky, 1996).
Lower limb velocities (Levanon & Dapena, 1988) are said to be an important determinant of ball velocity. Manolopoulos et al (2006) state that a greater shank velocity is indicative of a more “powerful” shot, the study conducted by Manolopoulous et al (2006) concluded that a strength and kick co-ordination training programme over a ten week period can cause an improvement in angular velocities of segments. A study conducted by Barfield et al (2002) found that a greater ball velocity was found with greater angular velocity of the distal segment (in male footballers). From this literature it can be assumed that a person with a high lower limb velocity should have a high ball velocity.
It has been theorised that the length, speed and angle of approach are the most important aspects of the preparatory phase, before movement transpires, having a significant effect on football kick success (Isokawa & Lees, 1988; Kellis et al, 2004).When a football kick is performed the athlete may kick the ball from a stationary position or approach the ball from a certain distance (Kathis & Kellis, 2007), Opavsky (1988) states that higher ball velocities are established when there is a running approach, of at least two to three steps, to the ball in contrast to a stagnant approach. Another important point is that a ball will in most cases be moving towards the player; consequently the player will not be hitting a stationary ball as is often the case in laboratory conditions, supported by Tol et al (2002). Kellis and Katis (2007) state that higher ball speed values have been during competition in contrast to a laboratory setting.
Isokawa & Lees (1988) concluded that on average maximum swing leg velocity occurred at an approach angle of between 30° and 45°, with a maximum velocity ensuing at 45°. From this finding it can be alleged that 45° is the optimal approach angle for a maximal velocity instep soccer kick (Clagg et al, 2009). Maximum ball speed and its relationship with accuracy is one which has been investigated with interesting results. Asami et al (1976) reported that by demanding both speed and accuracy from players, an 80% drop of the maximal value occurs as a result, this is a considerable reduction; however is further supported in literature stating that accurate kicking is achieved through slower ball velocities and kicking motion (Katis & Kellis, 2007; Lees & Nolan, 1998; Teixeira et al, 1999). Katis & Kellis (2007) deduce that a defined target, such as a goal, will determine the actual constraints on accuracy, with its manipulation leading to a trade-off between speed and accuracy of kick.
Another factor that could inhibit a maximal velocity IK is the kicking limb chosen. Many studies have found that higher ball velocities are found when football players kick with their dominant limb as opposed to kicks with the non-dominant limb; this has been attributed to higher foot speeds and a better inter-segmental pattern (Numone et al, 2006; Dorge et al, 2002); Manopoluous et al (2006) state that ball speed is the result of several segmental actions of the body during a kick, figure 3 illustrates the movements of the body segments during different phases of the kick.
Studies regarding male football performance in relation to kick biomechanics is a well researched area, however this does not correlate to the lack of knowledge gained when researching for female information. This statement is supported by McLean et al (2005) and Hewett et al (2006) who both acknowledge that few studies have characterised or examined female athletic performance in specific sports such as soccer, along with the assertion from Barfield (2002) that “the rapid rise in female participation in soccer worldwide has not been followed by a corresponding increase in the number of studies biomechanically that target female kicking patterns to determine if differences exist between males and females”. It is thought that the identification of kinematic differences between the sexes could potentially play a critical role in the teaching and training of aspiring female soccer players (Barfield et al, 2002). Consequently it can be assumed that female studies should be regarded to be of high importance and those found could help to eradicate huge differences between the sexes. With this said there are a few studies that have compared male and females, and studies that have solely looked at females.
A study by Barfield et al (2002) investigated differences between elite female and male soccer players. The study concluded that males kick the ball with greater ball velocity on the instep kick than women (see table 1 for mean ball velocity achieved in this study) and the differences in kinematic variables investigated were significantly different between the sexes, although this was small. However in this study there was one exception to the case, as it was found that one female generated greater ball velocity on two of her three kicks than the males on her dominant side, suggesting that not everyone follows the trend. A study by Tant et al (1991) supports Barfield et al (2002) findings, as it was found that male players produce greater ball speeds than their female counterparts, they attributed this finding to greater strength that males recorded; as tested on an isokinetic dynamometer.
In contrast to these findings, a study by Orloff et al (2008) comparing the kinetics and kinematics of the plant leg position between males and female collegiate soccer player during an instep kick, found that ball speed did not differ significantly between the two sexes as was hypothesised. Table 1 illustrates mean ball velocities, ranging from 15 to 30 m.s-1, achieved during a number of studies most of which occurred with the instep kick. Only one study shown provides details of a mean female ball velocity once more indicating the lack of research on female football participation.
It has been stated that kicking performance when measured by means of maximal ball velocity, can be improved by strength training (DeProft et al, 1988; Jelusic et al, 1992; Taiana et al, 1993), relating to Wisloff et al (2004) who states that maximal strength is an important factor in successful soccer performance; this is because of the apparent demands visible from the game. Strength has been defined as the ‘integrated results of several force producing muscles performing maximally, either isometrically or dynamically during a single voluntary effort of a defined task’ (Hoff & Helgerud, 2004); Schmidtbleicher (1992) states that strength influences all other components and thus it is located in an upper hierarchal level. The use of strength training is a common means of improving muscle function and has been said to develop performance of kicking skill through apposite training (Masuda et al, 2005).
Gomez et al (2008) believe that the coalescing of strength training with technical training involving motor tasks is required for improvements in performances to occur, this relates to the traditional training principle of specificity; Behm & Sale (1993) and Sale (1992) support this principle as they believe that training is intended to correspond to specificity in sport itself, this is in terms of contraction type, contraction force, movements and velocity. This can be related to football training, since the fundamental aspect of football is kicking and this involves a complex series of synergistic movements of the lower limbs, which in essence would be extremely complex to imitate with simple strength-training movements (Bangsbo, 1994).
Therefore strength training should be integrated into football training with several types and speeds of training involving the actual movement pattern in order to increase performance (Masuda et al, 2005). If a relationship between muscle strength and performance exists then it can be assumed that positive effects should become perceptible when measuring ball velocity, if these performance enhancing training benefits are not evident then athletes may not be motivated to participate in strength training (Myer et al, 2005).
Myer et al (2005) conducted a study that explored the effects that a comprehensive neuromuscular training programme had over a period of six weeks. The investigators measured performance and lower extremity movement biomechanics in female athletes, it was concluded that female athletes who trained with this six week programme could gain performance enhancements and significant improvements in movement biomechanics. Myer et al (2005) states that female athletes may especially benefit from multi-component neuromuscular training, as females often display decreased baseline levels of strength and power when compared with their male counterparts. The previous statement is supported by Kraemer et al (2003) and Kraemer et al (2001) who believe that a comprehensive training programme may significantly increase power, strength and neuromuscular control and therefore decrease gender differences in these measures.
Campo et al (2009) conducted a study over a period of 12 weeks on female soccer players; this involved the undertaking of a plyometric program. It was found that this program produced improvements in explosive strength in the female athletes and consequently this improvement could be transferred to soccer kick performance in terms of ball velocity; this study also lends evidence to the use of plyometrics in a strength training program.
Studies by Aagaard et al (1996) and Trolle et al (1993) found similarities within their results, since no significant improvements in kicking performance were established after knee-extension strength training. However De Proft et al (1988), Gomez et al (2008) and Monolopoulos et al (2006) all conducted strength training programmes that combined strength with another form of training, football training, plyometric exercises and technique exercises (kick co-ordination) respectively, found significant improvements in kicking performance (maximal instep football kick). The studies by Gomez et al (2008) and Myer et al (2005) took place over a 6wk period, with the study by Monolopoulos et al (2006) taking place over 10 weeks and Campo et al (2009) over a 12 week period, suggesting that the length of a training programme is interchangeable to gain relevant results. Hoff & Helgerud (2004) state that research based on strength training is often not conclusive; this may be due to the variances in measurement techniques.
Various studies have examined the muscle activation patterns that arise during a football kick; one of the findings to come from studies is the high activation of knee muscle groups (De Proft et al, 1988). To examine this further, maximal isokinetic data has been undertaken to study the moment of force of the knee extensors and flexors, this has been investigated in male players (Brady et al, 1993; Oberg et al, 1984; Oberg et al, 1986), female players (Reilly & Drust, 1997) and in relation to football kick performance (Cabri et al, 1988; Poulmedis, 1988; Reilly & Drust, 1997).
Rapid knee flexion and extension is an important part of a football kick as the knee flexes then extends at impact, this movement is accompanied by a stretch of the knee musculature during backswing ensued by immediate shortening during distal segment movement (Katis & Kellis, 2009). The action of the proximal segment being brought forward whilst the distal segment lags behind (as the knee is still flexing) serves to stretch the extensor muscles of the proximal segment before shortening of them is needed, this necessitates the generation of large end-point speed (Lees & Nolan, 1998). It can be assumed that if the knee extensor muscles are powerful then they should facilitate in large end point speed (greater ball velocity).
Isokinetic muscle testing is often used to evaluate strength within sport, with a range of data obtainable from its use (Ozcakar et al, 2003) however controversy surrounds its application. Wisloff et al (2004) believe that isokinetic tests do not reflect the actual movements of the lower limb segments during a football kick, and Dvir (1996) states that this is due to the nature of testing knee extensors, as it is a single-joint configuration, it is limited in functional scope.
A study by Reilly & Drust (1994), have reported results for female soccer players that show a high correlation between ball speed and knee extensor strength, this is supported by McLean and Tumilty (1993) who state that maximal strength of knee extensor muscles is an important determinant of kick performance. Asami et al (1982) report that the ball velocity and knee extensor strength relationship of the kicking limb may well depend on the skill level of the players, suggesting that the strength of the muscles in the knee has less input on ball velocity in football players whom are more skilled. This statement implies that less skilled players rely more on their muscular strength than skill. De Proft et al (1988) conducted a strength training programme for footballers and found a 25% increase in concentric muscle strength of extensors.
Studies have shown that knee extensor strength and kick performance however did not have a positive relationship, as for example Masuda et al (2005) found that knee extension/flexion strength was not correlated with the ball velocity and Aagaard et al (1996) conducted a 12 week training programme on the isokinetic strength of the knee extensors and flexors, with an increase in isokinetic and concentric strength found, but it was concluded that this gain did not help facilitate improvements in performance.
From current literature it is expected that the use of a strength training programme integrated with technical game play, will have a positive significant improvement from pre-test to post-test on both females and males as previous research has shown that a strength training programme improves performance (Manolopoulos et al., 2004; De Proft et al., 1988; Dutta & Subramanium, 2002), however the female group are expected to have a bigger improvement as they often have lower level of strength to start (Myer et al, 2005) leaving more room for improvement, and men will have a better kicking performance determined by ball velocity as they possess more power (Barfield et al, 2002; Tant et al, 1991). It is also expected that an improvement in knee muscle strength, limb velocity and foot velocity at ball contact will lead to an improvement in ball velocity as it can be said that kicking performance can be related to leg muscle strength as it is the muscles which are directly responsible for the increasing speed of the foot and therefore resultant ball velocity (Lees & Nolan, 1998). This information leads to the hypotheses for this study.
It is hypothesised that after a strength training and kick co-ordination programme both men and women will find significant improvements in their kicking performance and knee extensor strength, females will have a greater improvement in the pre to post test results than their male counterparts, men will have greater ball velocity both pre and post test than women, improvements in knee strength, limb velocity and foot velocity will lead to an improvement in ball velocity.
Before any real data collection commenced two pilot tests were conducted. This was to enable any aspects of the testing procedure to be checked, allowing areas of weakness and uncertainty to be enhanced and/or changes needing to occur to be implemented before actual testing transpired.
The first pilot test involved kinematic analysis data collection, using Qualysis Oqus 3D motion capture system, at a sampling rate of 500 Hz, under laboratory conditions. A participant was marked up with a lower limb marker set (see figure 7 and 8), a warm up and relevant instructions were given. 5 maximum velocity kicks were performed with the dominant foot at a target (1.82m x 1.2m) set 6 metres away from the position of the ball, a 2metre approach distance of self selected approach angle was allowed and a Sports radar precision gun (SRA 3000) was positioned behind the target. Uncertainties regarding target size, approach distance and quality of data collection were put under scrutiny. Collaboration with the participant allowed for uncertainties such as target size and distance of approach to be modified. Quality of data was checked and it appeared not all parts of the movement were captured or markers visible at all times (see figure 4). Due to these findings the pilot testing resulted in changes to the planned protocol, such as approach distance (an extra metre allowance was given), bandage size (was halved to prevent covering of markers), calibration technique in regards to area dynamically covered was increased (to cover all movement performed) and appropriate marker placement took place (incorrect palpitation had previously taken place).
The second pilot test was an extension of the first, relevant changes were made as noted in pilot test 1, with testing on the isokinetic dynamometer (ISOCOM- isokinetic technology, eurokinetics) included for strength data. A warm up was conducted prior to use, with the involvement of dynamic movements to help replicate the movement on the isokinetic dynamometer, once completed 5 practice trials took place followed by 3 trials that were collected as the data. This allowed for any time restraints for the two conjoined to be noted. It was found that the testing on the isocom took longer than the kinematic data, as it was expected that this would be roughly the same time; so implementation of a suitable time system could occur. Marker issues previously noted in pilot 1 were not problematic; this could be due to the increased area of calibration and anatomical landmark markers not being covered by bandages. Figure 5 shows that most data was captured and tracked, giving evidence to improvements made being effective, when comparing figure 4 and 5 against each other. The extra metre approach distance allowance proved successful with collection of data running more smoothly.
Sixteen amateur football players volunteered to participate in this study, eight females and eight males. Participants were split with regards to gender and assigned to either the female control group (FCG) (n= 4 females; age 20 ± 0.8 years; height 169 ± 5.8cm; body mass 68.9 ± 11.1 kg; all mean ± std), the male control group (MCG) ( n= 4 males; age, 21 ± 1 year; height 177.5 ± 7 cm; body mass 77 ± 10 kg; all mean ± std), the female strength training experimental group (FTG) (n= 4 females; age 20 ± 1.3 year; height 160 ± 1.8 cm; body mass 58.1 ± 4.3 kg; all mean ± std ) or the male strength training experimental group (MTG) (n= 4 males; age 17 ± 1. 2 year; height 174.9 ± 4.1 cm; body mass 73.1 ± 13.7 kg; all mean ± std). All females were right foot dominant, with 6 males being right foot dominant and 2 left foot dominant. Foot dominance was self selected based on the players’ reply to which foot they preferred kicking with to gain a maximal ball velocity outcome. Subjects were informed about the requirements, benefits and risks of the study, and completed an informed consent form and Par- Q prior to any testing (see Appendix).
In accordance to the study of Masuda et al (2004) kick performance was evaluated by measuring the maximal and mean velocity of the ball, by the use of Sports radar precision gun (SRA 3000), and a set number of trials in which the ball hit the target (5 times). Other measurements from this maximal kick were also taken by means of Qualysis Oqus 3D motion capture system, this uses multiple cameras (an eight camera system) to reconstruct three dimensional movement data; this was captured at a sampling rate of 500 Hz for 5 seconds. This enabled human movement analysis during the execution of a motor task (instep kick) to be traced via the use of reflective markers, gathering quantitative information (Cappozzo et al, 2005). The calibrated anatomical system technique (CAST) marker set was used (Cappozzo et al, 1995), each participant was instrumented with 44 retroflective markers (see figure 7 and 8). All markers and clusters used when capturing the data was with the aim to: “not significantly modify the performance being captured and measured” as stated by Brand and Crownshield (1981). These markers were placed on anatomical landmarks (medial and lateral side of proximal and distal end of segements) by palpation using guidance from Croce et al (2005), and on segments using clusters in accordance with Manal et al (2000) who state that a rigid shell with a cluster of four markers is the optimal configuration for a cluster set. Specific shoes were provided for all participants, with the anatomical landmarks placed on these prior to testing (figure 6).
Before data collection of each participant commenced, the motion capture system was calibrated (see figure 9a) to allow information gathering of spatial location of anatomical landmarks in regards to a known frame of reference (Cappozzo et al, 1994). This occurred by the use of a wand, which carries two markers of a given distance (750.5mm), and is made to coincide with the target anatomical landmarks by moving dynamically through the volume of cameras (Cappozzo et al, 2005; Richards, 2008) over a calibration frame placed in the data collection area (see figure 9b), this is removed so data for the intended activity (instep kick) can be performed and recorded. A static of each participant with clusters and anatomical markers was then taken (see figure 10) asking participants to gain a posture where anatomical markers can be seen by two or more cameras for at least a frame. Once this was achieved only tracking markers were kept on (thigh and shank clusters, anterior superior iliac spine (ASIS), posterior superior iliac spine (PSIS), greater trochanter, iliacs, foot markers except 1st and 5th metatarsals), as Cappozzo (1984) states markers used for identifying anatomical landmarks should be removed before physical movement is performed. Qualysis Track Manager (QTM) was the software used to capture the data including statics, dynamic movement and calibration.
The set-up design for the kick performance test can be seen in figure 11. Specific instructions were given to participants regarding their kicks, it was stated that although the kicks that missed the target would be repeated, they should not sacrifice speed in order to improve accuracy. A ball of standard size and standard inflation (Fédération Internationale de Football Association, FIFA, standard) was used. A warm-up was conducted, this took place on a treadmill (5-10 mins) followed by stretching, once the candidate felt they had been sufficiently warmed up practice trials took place. 2-3 practice trials were implemented allowing participants to acquaint themselves with test equipment and kicking conditions. Participants were allowed to self select their approach angle to the ball (between 0° to 60°), the only restraint utilised was the approach distance to the ball; participants were allowed a run up of between 2 to 3 metres this distance was marked and made noticeable to the participants. 5 successful trials were recorded on the dominant leg, a successful trial was classed as one which hit the target and the motion capture data was seen to be adequate, only 3 of these trials were analysed (data deemed as poor quality was discarded). This testing took place both pre and post intervention.
Once data collection had been completed in QTM, the information was used and markers labelled; including both static and dynamic data. For dynamic data, this took place through naming the markers in a dynamic frame and processing this to the corresponding data. An aim model was built, this was then generated and batch processed to all the dynamic trials, the checking of each anatomical frame ensued to ensure all markers were labelled correctly. Once this procedure was completed data was then exported to Visual 3D for further analysis and model building. Data from QTM (see figure 10) was built in to actual body segments that could be visually seen and recognised; this occurred through model building on Visual 3D. Figure 12 shows some examples of how the right hand side of the body was built; the same was done for the left side. Once model building had been completed, all trials were checked and different pipelines were put in place (a set of commands that can change or produce data wanted). An interpolation pipeline was conducted on the data to fill in missing data points, a ten frame gap fill was instrumented, filling in gaps more than this suggest that data is of poor quality. A low pass filter (using Butterworth filter) pipeline was put in place on the data, to smooth and remove noise that could be due to relative and absolute errors (soft tissue artefacts) (Richards, 2008), with a cut of frequency of 6 Hz used. Cut off frequencies previously used in other literature are between 6-18 Hz (Andersen et al, 1999; Dorge et al, 2002; Nunome et al, 2002; Teixeira, 1999). To determine heel strike of the non-kicking leg at placement the event minimum pipeline was used on the non-dominant leg (heel), to find the lowest point of the heel in the z axis (see figure 13). For information between a range of movements to be determined, the event ball contact was defined (see figure 14). Segment velocity (in x axis) of the thigh, shank and foot was extracted from the data, in the reports section, using the range of events previously defined (non-dominant leg heel contact and ball impact) to visually see data between and up to those chosen points.
Isokinetic concentric peak torque of the dominant leg was measured using an isokinetic dynamometer (ISOCOM- isokinetic technology, eurokinetics) see figure 15. The strength test involved movement of the knee (extension and flexion) to detect muscular strength in the knee extensor muscle groups. The angular velocity used for the movement was 60° s-1; this angular velocity has been used by many investigators to evaluate knee muscular strength of football players (Kellis et al, 2001; Ergun et al, 2004; Dauty et al, 2002). Prior to undergoing the test a warm-up was conducted, consisting of a 10 min warm up of cycling and 5mins of dynamic stretching, completion of this lead to the familiarisation process of the test protocols for the isokinetic movements that were tested including practice trials. Three maximum voluntary repetitions of flexion and extension at 60° s-1 took place in a seated position, with five familiarisation trials taking place beforehand, the participants were warned as to when the real trials were about to commence. The peak torque value was used to represent muscular strength; this is considered to be the gold standard in isokinetic measurement (ISOCOM testing and rehabilitation user manual). This testing took place both pre and post intervention with the same protocol applied for both testing.
The training programme undertaken in this study was a synthesis of findings derived from published articles for example training books and journals (Manolopoulous et al, 2006; Zatsiorsky & Kraemer, 2006; Chu, 1998). The FTG and MTG followed a 6-week training programme consisting of 1 session per week (each session consisting of up to an hour and a half) including a warm up and main activities. The main activity consisted of a circuit style fashion plyometrics, kick co-ordination and strength work ensemble, with exercises such as; lunges, squat jumps, resistance band work, core stability ball work and hurdle work included. Technical game play was incorporated into each session at the end of the circuit; with the aim of improving k