The development of an objective multi-dimensional approach to talent identification in junior Australian football

Talent identification (TID) is a pertinent component of the sports science discipline given the considerable influence it may have within the pursuit of excellence. Thus, research has attempted to identify the determinants of a talented performance through the use of objective testing procedures. However, many of these 'traditional' approaches have been operationalised by mono-dimensional objective physical performance tests that do not inherently account for the multi-dimensional requisites of game-play, particularly within a team sporting context. This is problematic when attempting to identify talent, as a successful performance in team sports is often the combination of physical, technical and tactical elements. For example, a physically inferior junior may still succeed against their physically superior counterparts given additional technical and tactical skills; commonly referred to as a compensation phenomenon. Hence, forecasting longitudinal performance based upon one element of effective play (e.g. physical) will likely lead to an unsubstantiated and biased identification. Despite the aforementioned, TID practices in junior Australian football (AF) are predominately facilitated by physically biased objective performance tests. Given the combative nature of game-play, physicality is an important attribute, but solely basing identification and selection on isolated physical attributes can be misleading given the previously mentioned compensation phenomenon. This mono-dimensionality is somewhat expected as to date there is a scarcity of objective tests measuring the multidimensional characteristics of AF game-play. Thus, through the consolidation of a number of theoretical concepts and recommendations proposed within the literature, this thesis aimed to develop a multi-dimensional objective approach to TID in junior AF, and in doing so, identify the determinants of a talented performance. To address this aim, objective physical, technical and tactical measurements were taken on both talent identified and non-talent identified junior AF players through the use of representative performance tests. Indeed, this reflected the first stage of the Expert Performance Approach (Ericsson & Williams, 1991; Williams & Ericsson, 2005) and the Model of a Skilful Player (Launder, 2001). Throughout each research study, talent identified players were defined through participation within the West Australian Football League (WAFL) State Under 18 (U18) Academy (an elite talent development program), whilst non-talent identified players were randomly chosen from the remaining cohort of WAFL U18 players not participating in the State Academy program. Thus, a cross-sectional observational research design was employed for each experimental procedure used throughout this thesis. It is of note that the first three studies utilised players from the 2013 sample, whilst the fourth research study utilised players from the 2014 sample. In the first of four research studies, a range of sport specific physical characteristics were found to differ between talent identified and non-talent identified junior AF players. However, a binary logistic regression model indicated that it was the measurements of standing height, lower body power and maximal aerobic capacity that provided the greatest prediction of talent, and thus important physical determinants of talent in AF at an U18 level. The second study investigated if measurements of technical skill could be used to accurately identify talent in junior AF. Despite the range of technical skills required in AF, the two modes of ball disposal (kicking and handballing) have been deemed critical for success based upon recent research (Parrington, Ball, MacMahon, 2013; Sullivan et al., 2014). Consequently, two representative skill tests were described; the Australian Football Kicking (AFK) test and the Australian Football Handballing (AFHB) test. Results indicated that the majority of the talent identified players possessed superior ball disposal skills in comparison to their non-talent identified counterparts. Specifically, measures of accuracy and ball speed on both the dominant and non-dominant sides reflected the strongest prediction of talent for the AFK test, and measures of accuracy on both dominant and non-dominant sides reflected the strongest prediction of talent for the AFHB test. These results reinforced the construct of each test, and highlighted their effectiveness for use as an objective TID tool in AF. Research had yet to investigate if decision-making skill was predictive of talent in junior AF despite its suggested importance for the exhibition of an expert performance in the game. The third study in this research series attempted to fill this remaining gap and objectively quantify decision-making skill through the use of a video-based decision-making task. In order to construct such a task, video footage was obtained from the Australian Football League (AFL) using an aerial behind-the-goal camera perspective. Through the use of an expert coaching panel, 26 clips out of an initial sample of 52 were deemed applicable, as each consisted of approximately three to five possible decision-making options. Results indicated that the talent identified players performed the task more accurately in comparison to their non-talent identified counterparts, and was thus a valuable objective tool for identifying talent at an U18 level. The fourth and final study in this research series investigated if the application of a multi-dimensional battery of objective performance tests provided more accurate TID in AF when compared to isolated performance measures. The construction of this test battery was informed by the results of studies one, two and three, but to ensure the translation of this test battery, it was applied to the 2014 U18 cohort, not the 2013 cohort which was done in the previous studies. However, the definition of talent identified and non-talent identified remained consistent with the previous studies. Results indicated that the majority of the talent identified players possessed a superior combination of physical, technical and tactical characteristics in comparison to their non-talent identified counterparts. Specifically, a receiver operating curve indicated a classification accuracy of 95% when summating the total scores obtained for each physical, technical and tactical test. This classification accuracy supports the implementation of multi-dimensional objective designs over the traditional monodimensional designs when attempting to identify talent in team sporting contexts. This thesis was motivated by the need to enhance the accuracy and reliability of current TID practices in AF by developing an objective multi-dimensional approach. In doing so, it contributes an important body of research to the study of TID by providing a conceptually translatable means in which the development of such an approach can be undertaken in other team sports.