Department of Exercise and Sport Sciences Faculty of Science, University of Copenhagen Nørre Allé 51, DK-2200 Copenhagen N Tel.: +45 3532 0829 Fax: +45 3532 0870 E-mail: ifi@ifi.ku.dk – www.ifi.ku.dk
Talent Development in Danish Elite Athletes
A controversial question within elite sports is whether or not young athletes need to specialize at an early age, or if it is more beneficial to follow the path of early diversification. This path includes sampling different sport experiences during childhood and then specializing later during adolescence. To explore this question, the career paths of Danish elite athletes were investigated. The main research question addressed the differences between elite and near-elite athletes using data concerning the amount of practice hours during the career, engagement in additional sports, time of specialization into the main sport, as well as the sport-specific achievement motive and volitional factors. A total of 722 Danish elite athletes from 34 different sports replied to the questionnaire. In order to prevent a too heterogeneous sample, all analyses were conducted for groups of sports with similar requirements. The results concerning the career paths of athletes from cgs sports, team sports, precision sports, and racquet sports are presented and discussed. Moreover, findings on the differences between elite, near-elite athletes, and dropouts are provided for cgs athletes and football players.
Karin Moesch Anne-Marie Elbe Marie-Louise Trier Hauge Johan Wikman
Talent Development in Danish Elite Athletes Report for the project financed by Team Danmark, 1/5/2009 – 30/9/2010
Karin Moesch, Anne‐Marie Elbe, Marie‐Louise Trier Hauge and Johan Wikman Institut for Idræt Københavns Universitet 2011
Talent Development in Danish Elite Athletes Report for the project financed by Team Danmark, 1/5/2009 – 30/9/2010 © Karin Moesch, Anne‐Marie Elbe, Marie‐Louise Trier Hauge and Johan Wikman Department of Exercise and Sport Sciences, University of Copenhagen 2011 Front page layout: Allis Skovbjerg Jepsen Photos: Das Büro and Team Danmark. Layout: Marie‐Louise Trier Hauge Print: Det Samfundsvidenskabelige Fakultets ReproCenter
Talent Development in Danish Elite Athletes
Contents 1. Introduction and theoretical background ........................................................................................ 5 Elite Performance through Early Specialization ......................................................................... 5 Elite Performance through Early Diversification ........................................................................ 7 Career Development Stages .......................................................................................................... 8 2. Aim of the project .............................................................................................................................. 9 3. Method .............................................................................................................................................. 10 Design ........................................................................................................................................... 10 Procedure ..................................................................................................................................... 10 Sample .......................................................................................................................................... 11 Instruments .................................................................................................................................. 14 Data analyses ................................................................................................................................ 16 4. Results, discussion and practical implications of the different sport categories ....................... 17 Addendum I: Validation of data about practice hours .............................................................. 17 4.1. Cgs Sports .................................................................................................................................. 17 Sample Cgs Sports ........................................................................................................................ 17 Results Cgs Sports ........................................................................................................................ 18 Discussion Cgs sports .................................................................................................................. 19 Practical implications Cgs sports ................................................................................................ 21 4.2. Team Sports .............................................................................................................................. 23 Sample Team Sports .................................................................................................................... 23 Results Team Sports .................................................................................................................... 24 Discussion Team sports .............................................................................................................. 25 Practical implications Team Sports ............................................................................................ 29 3
Talent Development in Danish Elite Athletes
4.3 Precision Sports ......................................................................................................................... 30 Sample Precision Sports .............................................................................................................. 30 Results Precision Sports, ............................................................................................................. 30 4.4 Racket Sports, ............................................................................................................................. 31 Sample Racket Sports .................................................................................................................. 31 Results Racket Sports .................................................................................................................. 32 Addendum II: Limitations of the study ...................................................................................... 32 References ............................................................................................................................................ 34 Appendixes ........................................................................................................................................... 38 Appendix 1: Cgs Sports ................................................................................................................ 38 Appendix 2: Cgs Sports ................................................................................................................ 39 Appendix 3: Team Sports ............................................................................................................ 40 Appendix 4: Team Sports ‐ football ............................................................................................ 41 Appendix 5: Team Sports ............................................................................................................ 42 Appendix 6: Precision Sports ...................................................................................................... 43 Appendix 7: Racket Sports ......................................................................................................... 44
4
Talent Development in Danish Elite Athletes
1. Introduction and theoretical background The question of how to achieve peak performance is central in elite sports. Researchers within all domains of sport sciences hope to gain knowledge on which variables and processes lead to winning international medals. Within social sciences, and from a developmental perspective, one of the controversial questions concerns which career path leads to expert performance. Based on the “Developmental Model of Sport Participation” (Côte, Baker & Abernethy, 2007), two ways to reach elite performance are described. The path of early specialization focuses on early involvement in the main sport, normally occurring in early to middle childhood, with very little or no involvement in other sports. The importance of a high amount of deliberate practice, defined as a highly‐structured and goal‐oriented activity aimed at improving the current level of performance, is stressed during all ages (Ericsson, Krampe & Tesch Römer, 1993). Additionally, emphasis is placed on constraint factors, including motivation and effort, which are considered essential to maintaining the hard and sometimes monotonous training regime. In contrast, the path of early diversification postulates that the first years of sport participation should be characterized by the involvement in different sports, as well as a high amount of play‐like practice that focuses little on deliberate practice activities. Following these sampling years, around age 12, the young athlete gradually reduces his/her involvement in other sports and shifts focus to the main sport, beginning a highly‐deliberate practice regime around age 16 (Côté, Baker & Abernethy, 2007). The next sections will describe the two paths in detail.
Elite Performance through Early Specialization Emerging from Ericsson et al.´s (1993) theoretical framework, this path postulates that in order to achieve expertise, one must engage in 10,000 hours of deliberate practice within the chosen domain. The theory is based on a well‐documented, strong and positive relationship between amount of practice hours and performance found in different domains (e.g. Ericsson et al., 1993). Ericsson et al. (1993) also argue that the accumulation of these practice hours must correspond with sensitive stages of the biological and cognitive development during childhood and adolescence. A logical conclusion of the paradigm suggests that an early start in a given sport is a necessary requirement to reach expertise and that not doing so will result in a practice delay compared to peers who started their sport involvement earlier. There is extensive scientific evidence from different sports that supports a positive relationship between practice hours and expertise level (e.g. Baker, Côte & Deakin, 2005; Baker, Deakin & Côte, 2005; Helsen, Starkes & Hodges, 1998; Hodges & Starkes, 1996; Hodges, Kerr, Starkes, Weir & Nannanidou, 2004; Law, Côte & Ericsson, 2007). In order to persevere on the 5
Talent Development in Danish Elite Athletes
long and strenuous path to expertise, including deliberate practice that are not considered inherently enjoyable, Ericsson et al. (1993) suggest three domains to be essential in developing expertise. Aside from resource constraints (e.g. access to training facilities and coaches or parental support) that assumingly play a crucial role in the development of elite sport performance (Holt & Dunn, 2004; Van Yperen, 2009; Baker & Horton, 2004), Ericsson et al.´s (1993) focus is on motivation and effort. The motivational constraint refers to an individual’s goal commitment. The effort constraint refers to the ability of an individual to persist in high amounts of deliberate practice; this constraint is comparable to the concept of volition, as discussed in the Rubicon model of action phases (Heckhausen, 1989). This model stresses the assumption that motivation needs to be complemented by volition or will‐strength in order for an intention to be transformed into an action. In other words, motivation alone is not sufficient to maintain athletic training over the long period of time required to achieve expertise. Motivation needs to be reinforced with volitional processes that are responsible for initiating an action, despite internal and external resistance, and for maintaining that action until the goal has been reached (Kuhl, 1983). Several studies confirm the significant role that motivational and volitional factors play in the involvement and performance level of elite‐sport athletes (e.g. Beckmann & Kazén, 1994; Elbe, Beckmann & Szymanski, 2003; Holt & Dunn, 2004; Van Yperen, 2009; Wenhold, Elbe & Beckmann, 2009). Even though the relationship between practice and performance is one of the most robust in behavioral science (Baker, Deakin & Côte, 2005), criticism arose regarding Ericsson et al.’s (1993) approach. Firstly, even though many studies revealed that elite performers trained more than sub‐elite performers, the elite performers failed to reach the magic number of 10,000 practice hours (Van Rossum, 2000; Baker, Côté & Abernethy, 2003). Secondly, Baker and Côté (2006) reveal that reducing the development of expertise in sport to simply deliberate practice fails to acknowledge important developmental, psycho‐social, and motivational factors of young athletes. Thirdly, there is no consensus stating that early onset and early specialization are required for the development of expertise (e.g. Carlson, 1988; Barynina & Vaitsekhovskii, 1992; Lidor & Lavyan, 2002). For example, the results of Vaeyens, Güllich, War and Phillippaerts (2009) indicate that there is no evidence that an early onset and a higher amount of sport‐ specific training are associated with greater success at a later stage. Additionally, a body of research emerged showing that early specialization can lead to negative consequences for the athletes, such as attrition and negative health outcomes (e.g. Côté, Baker & Abernethy, 2007). Law, Côté and Ericsson (2007) found that Olympic‐level rhythmic gymnasts, who had acquired significantly more training hours in their career than their international‐level peers, rated their health as lower and their participation experiences as less fun. Gould, Udry, Tuffey and Loehr’s (1996) study revealed that early specialization and highly‐ structured training reduced intrinsic motivation and led to higher dropout and burnout rates 6
Talent Development in Danish Elite Athletes
among young athletes. Likewise, Wall and Côté (2007) found that athletes who dropped out of sport, compared with athletes who continued their participation, had began off‐ice training earlier in their careers. This indicates that early specialized training regimes that are not inherently enjoyable can have a detrimental effect on the long‐term development of athletic expertise. These results strengthen the assumption that in order to become a highly motivated, self‐determined, and committed adult athlete, it is crucial to build a solid foundation of intrinsic motivation at early stages (Deci & Ryan, 2000). No individual involved in elite sports will negate deliberate practice as an important pillar for reaching expertise, and the prominence of practice is generally agreed upon in literature (Janelle & Hillman, 2003). However, the risks of an early and intense involvement in sports as well as the evidence for late specializing experts need to be acknowledged. Therefore, it has to be questioned whether or not early specialization is the exclusive path to expertise. It also needs to be investigated if different paths that involve lower risks for the individual, can lead to the same outcome (Baker, Coté & Deakin, 2005).
Elite Performance through Early Diversification Based on the above‐mentioned results, the notion emerged that, in addition to early specialization, expertise can be reached through early diversification (Côté, Baker & Abernethy, 2007). Two underlying notions exist for that path. From a psycho‐social point of view, it can be reasoned that engaging in a variety of different sports allows the young athlete to experience different physical, cognitive, affective, and psycho‐social environments (Côté, Lidor & Hackfort, 2009). It is hypothesised that this path promotes the development of intrinsic motivation (Côté et al., 2007), which again serves as a basis for a self‐regulated involvement in elite sport at a later stage (Côté et al., 2009). From a performance point of view, it can be argued that experiences in various environments provide the young athlete with important physical, personal, and mental skills required to specialize in one sport at a later stage in his/her career (Côté et al., 2009). The central notion of performance point of view is that motor, cardio‐ vascular, and mental skills can be transferred from one domain to another. Even with limited scientific research (Feltovich, Prietula & Ericsson, 2006), there remains a general assumption that talented athletes can transfer common skills across sports (Williams & Ford, 2008). Moreover, current research suggests that the effect of skills transfer is most pronounced during early stages of involvement (Schmidt & Wrisberg, 2000), corresponding with the timeframe of the sampling years in the “Developmental Model of Sport Participation” (Côté et al., 2007). Evidence shows that later specialization can prove more beneficial while training to become an expert athlete. Carlson (1988) found that elite tennis players specialized later and 7
Talent Development in Danish Elite Athletes
practiced less than their sub‐elite peers between the ages of 13 and 15, but intensified their training considerably more after age 15. Likewise, Lidor and Lavyan (2002) found that elite athletes from various sports began specialisation later than sub‐elite athletes. Nevertheless, the elite athletes had completed more training hours by the time they reached peak performance, indicating that despite their late start, they still managed to compile ample hours to perform at the top level. Barynina and Vaitsekhovskii (1992) found that swimmers who specialized early, when compared with swimmers who specialized later, spent less time on the national team and ended their sport career earlier. Güllich’s (2007) results showed that early intensification in athletic development does not correlate with long‐term success, but that in contrast, particularly successful careers are characterized by a deceleration of practice and competitive development. Lidor and Lavyan´s result (2002) confirms the idea of sampling, finding that 70% of the elite, compared to 58% of the sub‐elite athletes, performed more than one sport in their early years of involvement. Likewise, Emrich and Güllich (2005) report that both, being active in another sport besides the main sport as well as starting the sport career in another sport and then switching to the main sport at a later age, are significantly more prevalent in German athletes who were successful at the international level compared to their peers who competed at only the national level. Evidence suggests a beneficial effect of early diversification, not only on performance level, but also on other variables. Baker and Côté (2006) state that sampling and deliberate play in the early years of sport participation may lead to more enjoyment and a lower frequency of dropout, which indirectly contributes to the attainment of a high level of performance in adult years. Moreover, they report that athletes who sample and diversify in their young years may be less at risk for injuries than their peers that specialize early. However, doubts arose concerning whether or not sampling is inherently beneficial for all young athletes; in particular, several authors questioned the application of early diversification to all sports (Baker, 2003; Williams & Ford, 2008). Furthermore, Côté et al. (2009) conclude that early diversification is not beneficial for athletes in sports where peak performance occurs before full maturation, such as gymnastics. Emrich and Güllich’s (2005) study confirms this assumption.
Career Development Stages In addition to the above mentioned “Developmental Model of Sport Participation” (Côté et al., 2007), another approach to describe athletes’ career development exists. This approach takes into account the age at which athletes pass through different transitions. Based on Bloom’s (1985) stages of talent development, Wylleman and Lavallee (2004) designed a model that focuses on the athletic development, as well as the psychological, psycho‐social, and academic development of athletes. They describe three transitions which take place during a sport career: a transition into organized sport (entering initiation stage), a transition to a more intense level 8
Talent Development in Danish Elite Athletes
of training and competition (entering developmental stage), and a transition into the elite level (entering perfection stage). Along with suggesting timeframes in which athletes typically go through these transitions, Wylleman and Lavallee (2004) also stress that there are sport specific differences that should be taken into account when investigating career development.
2. Aim of the project Currently there is no quantitative data concerning the career development of Danish elite athletes available. In attempt to bridge this gap, the aim of this project seeks to gather and compare data on the careers of Danish athletes from different levels. The main research question addresses differences between elite and sub‐elite athletes within the following areas:
the amount of practice hours they sample during their career
their engagement in additional sports during their career
the time point of their specialization into the main sport
the sport‐specific achievement motive
volitional factors
Moreover, for the sport type categories that had an ample sample size (cgs and team), logistic regressions were performed to investigate which of the above mentioned variables predict membership in the elite group. An additional research question was investigated in hopes of detecting differences between elite, sub‐elite, and dropout athletes regarding their engagement in other sports, the time point of their specialization into the main sport, as well as the sport‐specific achievement motivation and volitional factors. Due to sample size, this research question was only investigated within the cgs group and the football players.
9
Talent Development in Danish Elite Athletes
3. Method Design In order to gain additional information concerning the optimal path for reaching high‐level athletic performance, it seems meaningful to identify variables that differentiate elite athletes from sub‐elite athletes based on exposure to practice activities (Williams & Ford, 2008). Many studies within the domain of talent development and expertise have been conducted based on the seminal work of Bloom (1985), using a retrospective design. Even though this design bears methodological risks (e.g. recall bias, see Hodges et al., 2007), it can provide interesting and meaningful insights into the early experiences of elite and sub‐elite athletes when there is not enough resources for longitudinal studies. Based on the above stated considerations, the present study adopts a cross‐sectional, retrospective design.
Procedure A link to a web‐based questionnaire was sent out to the target group by email. A web‐based design was chosen because it seemed most suitable for a sample involving young persons. Web‐ based studies offer the advantage that the participants can choose individually when they want to answer and are also a low‐cost method for obtaining responses from participants from different parts of the country (Shaugnessy et al., 2006). Prior to starting the questionnaire, the athletes were informed about the content and the aim of the research project, as well as being told that all the data would be treated confidentially and that participation was voluntary. After six weeks, a re‐test was sent out to the participating athletes with the aim of checking the validity of some of the variables. In order to increase response rates, reminders were sent out by mail and/or SMS after both surveys. To further check the data’s validity, some of the participants who simultaneously took part in an interview study conducted by another Danish research group were on that occasion asked the same questions again, offering the unique opportunity for another validation check four months after data collection.
Unfortunately, there were only a few athletes from the first data group that could be
categorised as dropouts (see a description of dropout below). Because of this small number, it was decided to conduct another data collection five months after the first one. E‐mail addresses from potential dropout athletes were collected through the contact of various federations, clubs, and coaches. All these athletes received an e‐mail with a link to the same questionnaire as the athletes from the first data collection phase. Moreover, an e‐mail with the link to the questionnaire was also sent to PE students at the Department of Sport and Exercise Sciences at the University of Copenhagen, asking for their participation in the study, assuming they met the
10
Talent Development in Danish Elite Athletes
criteria for dropouts. A reminder email was sent to all the participants two weeks after the first one was sent out.
Sample All athletes that were registered in Team
Sport
N
Sport
N
Athletics
43
Pentathlon
1
Automobile
2
Motorsport
15
Badminton
38
Orienteering
19
Basketball
1
Riding
6
Table tennis
12
Rowing
47
Bowling
9
Sailing
52
Wrestling
6
Ski
1
Archery
7
Shooting
7
Curling
13
Sport dance
2
Cycling
37
Squash
1
Football
124
Swimming
68
Golf
25
Taekwondo
13
Gymnastics
4
Tennis
7
were involved in 34 different sports. Table 1
Handicap sport
3
Triathlon
12
shows the distribution of athletes (who filled out
Handball
63
Waterskiing
1
the questionnaire to at least some extent) within
Ice hockey
45
Volleyball
9
the different sports.
Kano / Kayak
13
Others
16
Danmark’s database (Denmark’s elite sport organization), and who were supported in the year of the survey (2009) or had been supported within the last six years were contacted. From the initial 1,914 athletes, 743 replied (38,8%). 17 cases had to be deleted because they stopped answering after only a few questions, another 4 cases had to be deleted due to unreasonable answers, which left 722 athletes. 301 (41.7%) were female and 421 (58.3%) male, ranging from 13 to 53 years of age with an average age of 23.77 (SD = 6.81). Amongst these athletes, 538 were still involved in their main sport, while 185 had retired before the survey took place. The athletes
Table 1: Distribution of athletes within the different sports
Unfortunately, many athletes did not fill in the
questionnaire completely, resulting in a data file with numerous missing variables. Because of these gaps in the data, sport‐specific evaluations were not possible: reason being, if the sample size is too small, the resulting power would have been too low, and/or because some statistical analyses were not possible with small sample sizes at all. However, Emrich and Pitsch (1998) propose that sports sharing similar structural conditions should lead to similar career paths, which justifies analyzing such similar sports together. Other studies also followed that approach, analyzing data of athletes from different sports with similar structural exigencies (e.g. Güllich, 2007). Therefore, it was decided to group the sports into the following categories (table 2):
11
Talent Development in Danish Elite Athletes
Team sports: all sports that are performed in a team, opposing another team (e.g. football, team handball).
Cgs sports: individual sports that are measured in centimetres, grams or seconds (e.g. rowing, cycling).
Aesthetic sports: sports that are evaluated by external judges (e.g. gymnastics, sport dance). Racket sports are game sports executed with rackets (e.g. badminton, table tennis).
Combat sports: defined as individual sports where two opponents are, based on the respective rules, fighting against each other (e.g. taekwondo, wrestling).
Precision sports: sports performed in a team or individually, where precision is the decisive factor (e.g. golf, bowling, shooting).
Motor sports: sports performed with a motorised machine (e.g. motocross, speedway).
Others sports: contain sports that could not be assigned into one of the existing categories.
Sport category
N
%
Because only two categories included a sufficient
Team Sports
245
33.9
number of athletes for performing regression
Cgs sports
295
40.9
analyses, the focus of the evaluations of the project
Aesthetic sport
7
1.0
Racket sport
58
8.0
Combat sport
24
3.3
Precision sport
63
8.7
Motor sport
18
2.5
sport categories were not analysed more in‐depth
Other sports
12
1.7
due to the small number of athletes and the inability
Table 2: The number of athletes in the different sport categories.
was placed on team sports and cgs sports. The two next biggest categories (racket sports and precision sports) will be briefly addressed. However, the other
of doing statistical analyses with such small sample sizes.
The “elite” category (n = 295) was defined by a placement in the top 10 at a world level championship (e.g. World Cup, Olympics) or by winning a medal at a championship at the European level (e.g. European Championship) on a senior level. In order to eliminate an age bias, athletes up to age 21 were also categorised as elite if they had won a medal at a junior championship at a world level. All athletes who did not meet these criteria were labelled as sub‐ elite athletes (n = 275). Additionally, when dealing with the categorization of elite and sub‐elite athletes, missing answers posed a problem: 152 athletes out of the whole sample did not fill in the questions about sport success. Therefore, these athletes cannot be labelled as either elite or sub‐elite athletes and their data cannot be used in the analyses.
12
Talent Development in Danish Elite Athletes
From the second data collection (dropouts), 76 athletes completed the questionnaire. 2 questionnaires were deleted because they only filled in the first question, and the data of another 9 athletes was removed from the file because they answered that they were still involved in their main sport, therefore not qualifying as a dropout. This led to a total of 65 dropout athletes that were then added to the main file. Two researchers blindly categorized the athletes who indicated that they had stopped their engagement in their main sport into dropout or non‐dropout; this was done based on their reasoning for retirement (which was formulated as an open question). The cases in which the two researchers did not agree were discussed in the research group, and from there a decisive categorization occurred. Throughout the process, the following criteria served as a general guideline for the categorization of a dropout: a) Lack of motivation for sport engagement b) Performance results were not satisfying c) Missed an important qualification d) Educational / vocational reasons (started university, got a job offer, etc.) e) Lack of time for a high training regime f) Injuries categorized as not being serious enough for a career termination
g) Age: In general, athletes were only categorized as dropouts until the age of 20‐22. An age range was used because different sports have different ages of peak performance and therefore also different dropout ages.
As additional information, the answer to the following question, “at which time point during your career did you retire from
Sport category
N
sport” (answer possibilities: “When I retired, I had not yet
Cgs sports
52
reached my personal peak performance.”, “When I retired, I was
Team sports
25
at the peak of my performance.”, “When I retired, I had already
Aesthetic sports
3
passed my personal peak performance.”), was taken into
Racket sports
6
Precision sports
9
Total
95
account.
This procedure resulted in a total of 95 athletes being
categorized as a dropout. The distribution of the dropout athletes within the different sport categories is shown in table 3.
13
Table 3
Talent Development in Danish Elite Athletes
Because the number of dropout athletes in aesthetics sports, racket sports, and precision sports is too small for statistical analyses, it was decided to conduct the dropout analyses for only the cgs and team athletes. A closer look at the distribution of the dropouts in the team sport athletes showed that most of these athletes were involved in football (table 4). Therefore, it was
Sport
N
Basketball
1
Football
19
Handball
2
Ice hockey
1
Volleyball
1
decided to conduct those analyses sport specifically (football) by
Other (floorball)
1
comparing elite players with sub‐elite and dropouts.
Total
25
Table 4
Instruments The questionnaire covered information on the following topics: 1.
Biographical information
2.
Practice hours in the main sport: The athletes reported how many hours they trained on average per week for every year in their main sport, starting with the current year and then working backwards (see Hodges et al., 2007).
3.
Involvement in other sports
4.
Career development: The athletes stated the age they entered the “initiation stage1,” the “developmental stage2” and the “perfection stage3” (Wylleman & Lavallee, 2004), the age they participated in their first national, and international competition as well as how many years they were a member of the junior and senior national team.
5.
Weekly training schedule: For data validation purposes, the athletes reported their average training schedule for every weekday during the current year or, alternatively, for the last year they were involved in their main sport at an elite level.
6.
Athletic success: The athletes gave their results from different international competitions at the junior and senior levels.
1 Initiation stage starts when athletes first enter their sport in an organized setting (e.g. entering a club).
During this stage, athletes are engaged in fun, playful sport and perceive sport as merely playing a game.
2 During the developmental stage, the amount of training increases, athletes specialize in one sport and
start competing on a regional / national level. Typically during this stage, athletes narrow their focus to one or two sport disciplines that they are hooked by and committed to. 3 At the Perfection stage, athletes start competing at the highest level, at international competitions. During this stage, athletes become experts in their sport and feel responsible for their practices and competition performances.
14
Talent Development in Danish Elite Athletes
In order to gather data on the constraint factors, motivation and effort, it was decided to use the Rubicon model of action phases (Heckhausen, 1989; see Introduction) as a theoretical base,. This model appears to best represent the idea of motivation and effort that is proposed by Ericsson et al. (1993). The following measurement instruments were chose based on the availability of questionnaires in Danish as well as good reliability and validity in previous projects:
7.
The short version of the Achievement Motives Scale‐Sport (Elbe & Wenhold, 2005) assesses the two achievement motive components, hope for success and fear of failure. Each scale has 5 items, and uses a Likert‐scale answering format ranging from 0 (not true for me at all) to 3 (exactly true for me). The values for the scales range from 0 (very low) to 15 (very high). The two scales show high internal consistency with the current sample (Hope for success: Cronbach’s alpha4 = .83, N = 573; Fear of failure: Cronbach’s alpha = .85, N = 573).
8.
The Volitional Components Questionnaire Sport (VCQ‐Sport; Wenhold et al., 2009c) measures volitional skills and deficits in relationship to training and competitions. It assesses 60 items through 20 scales within 4 main components (self optimization, self impediment, lack of activation, and loss of focus). The questionnaire has a Likert‐scale answering format ranging from 0 (very low, “not true for me at all”) to 3 (very high, “exactly true for me”). The scales are formed by taking the average of all items, resulting in scale values ranging from 0 (very low) to 3 (very high). Due to the length of the questionnaire, the present study focuses on four scales: selfdetermination (Danish version: 4 items), lack of energy (4), postponing training (3) and avoiding effort (4). The scales were meaningful for the research question and showed good psychometric properties in the Danish version (Cronbach’s alpha between .68 and .83; Test‐retest reliability between .67 and .70; Wikman, 2007). The scales exhibited acceptable internal consistency for the present sample (lack of energy: Cronbach’s alpha = .71, N = 563; postponing training: Cronbach’s alpha = .78, N = 563; avoiding effort: Cronbach’s alpha = .68, N = 563; and self‐determination: Cronbach’s alpha = .61, N = 563).
4 The Cronbach’s alpha are based on the analyses of the complete sample of the project, involving the sample from team sports as well as athletes from other sport categories
15
Talent Development in Danish Elite Athletes
Data analyses As previously mentioned, missing data present a challenge in research. Due to the length of the question on practice hours in the main sport, the present study unfortunately revealed a high amount of missing values from that question. Since that information was the heart of the whole project, it was decided to not estimate the missing data. Outliers were detected and adapted to a more appropriate value based on the z‐value as well as through discussions within the author team according to the suggestions of Tabachnick and Fidell (2007).
After collecting data from the main survey and the two re‐tests, correlations were
performed to analyze the validity of the data on practice hours in the main sport. This was done as retrospective; data can be biased, therefore checking the data before analyzing seems indispensable.
In order to investigate differences between the elite and the sub‐elite samples, in terms
of the variables related to practice hours in the main sport, involvement in other sports, and data on career development, T‐Tests were conducted with a significance level of .05. For the categories that were big enough (cgs sports and team sports), additional analyses regarding predictions could be included. A logistic regression was performed to investigate whether practice hours in the main sport, involvement in other sports, data on career development, as well as motivational and volitional variables (independent variables) predicted membership in the elite athlete group (dependent variable). The enter method was chosen because there are no hypotheses concerning the order of importance of predictor variables. Assumptions regarding the distribution of the predictor variables are not required for logistic regressions (Tabachnick & Fidell, 2007).
16
Talent Development in Danish Elite Athletes
4. Results, discussion and practical implications of the different sport categories Addendum I: Validation of data about practice hours Three different measures were used to validate the data on practice hours. 1) A correlation was performed between two measures given in separate sections of the questionnaire, both intended to collect the same information (e.g. the amount of weekly training in the data on practice hours history and the information about the average training amount per week from the same year). The correlation between the two measures was .70 (N = 459). 2) The average result of the written re‐test (4 weeks after the data collection), over the seven different time points, was .75 for the weekly training amount. 3) The results of the re‐test (4 months after the data collection) collected during the interview study with 16 athletes, showed a correlation of .74 for the weekly training amount. All correlations can be categorized as strong (Brace et al., 2009). Additionally, analyses revealed that the correlations in the elite and the sub‐elite athletes (elite athletes: .76, sub‐elite athletes: .74) did not differ, indicating that the two groups have a similar level of recall. It can be concluded that the data of the present study is reliable and also comparable in quality to the data given from similarly structured studies.
4.1. Cgs Sports Sample Cgs Sports
Sport
Ntotal
nelite
nsubelite
Out of the 295 athletes involved in cgs sports, 243
Canoeing/kayak Cycling
12 34
11 28
1 6
Orienteering
17
6
11
Rowing
40
35
5
to the elite category and 95 to the sub‐elite. Table 5
Sailing
39
23
16
displays the distribution of sport and success level in
Skiing
1
1
0
the sample.
Swimming
55
24
31
Track and field
33
11
22
Triathlon
11
8
3
Weightlifting
1
0
1
Total
243
148
95
qualified as either elite or sub‐elite athletes and were therefore entered into the analyses. 148 athletes belong
161 athletes were currently active in their main
sport at an elite level, while 82 athletes retired before the survey was conducted. The mean age for the 96 female and 147 male athletes was 24.5 years (SD = 7.5), ranging from 13 to 51 years of age. The elite athletes were older (M = 26.58, SD = 7.49) than the sub‐elite athletes (M = 21.16, SD = 6.16).
17
Table 5: The distribution of sport and success levels in the sample.
Talent Development in Danish Elite Athletes
Results Cgs Sports T‐tests reveal significant differences between the elite and sub‐elite athletes in 11 of 26 variables (Appendix 1). Concerning the data on practice hours in the main sport, the results show that the sub‐elite athletes have completed significantly more training hours, some as early as age nine, and they continue to compile more hours throughout early adolescence, until age 15. The effect sizes are considered to be moderate (0.45 ≤ d ≤ 0.50; Cohen, 1969). At age 18, the amount of practice hours is roughly the same for the two groups. After age 18, the elite athletes complete more hours, resulting in a significant difference by age 21 from the sub‐elites, whose training increase has not developed that intensively. Figure 1 shows the development of practice hours at five different time intervals.
7000 6000 5000 4000 3000 2000
Elite
1000
Sub‐elite
0
Figure 1: Development of practice hours at five different time points.
However, elite and sub‐elite athletes do not differ in their involvement in other sports. Regarding different variables on career development, the following results can be found. Elite athletes state that they pass important steps within their career (e.g. starting sport, participation at first competition, etc.) at a significantly older age than the sub‐elite athletes (0.40 ≤ d ≤ 0.63). Moreover, the elite athletes spend significantly fewer years on the junior national team (d = 0.27), but more years on the senior national team (d = 0.97). No significant differences between the two groups could be found regarding motivational and volitional factors, indicating that the two groups do not show different characteristics within these domains. In a first logistic regression, six variables (membership on junior national team, membership on senior national team, age, and training up to age 12, 15 and 18) were
18
Talent Development in Danish Elite Athletes
significant, and were re‐entered in a second logistic regression. In this analysis, a total of 175 cases were analyzed, with the full model significantly predicting membership in the elite group (2 = 91.51, df = 6, p
