Background

An increasing number of talent identification and development (TID) programs are installed by sports associations to identify young athletes with the potential to achieve future success and provide them with the most optimal opportunities and environments to develop [1,2,3]. However, identifying talent and discovering how to optimally stimulate young athletes’ development remains a difficult challenge in most sports and contexts. Consequently, sports associations are searching for new solutions to increase a program’s efficacy and efficiency. One could argue that young athletes need to require a certain baseline performance/skill level to be able to compete and be eligible for a TID program in the first place. As such, programs include the development and implementation of assessments that measure certain individual characteristics (e.g., anthropometric, physiological, technical) suggested to influence or determine future performance. Unfortunately, it appears that most objective measures can only explain a small part of future performance. This might be due to the use of unidimensional and/or cross-sectional designs in most conventional TID studies. Such approaches seem to be oversimplifying the complexity and evolving nature of talent in sports [1]. It is uncontentious that athletic performance is multidimensional [1, 4] and that the developmental process is a highly individual and nonlinear pathway involving many interacting factors [3, 5,6,7,8,9,10,11]. As a response, scientists have been calling for studies using multidimensional and longitudinal approaches to better understand the complex nature of talent [12].

The Groninger Sports and Talent Model (GSTM) for the development of a talented athlete’s performance is in accordance with such an approach as it suggests defining performance in a multidimensional way and monitoring development over time [13, 14]. The model is based on Newell’s constraints-led model and defines sports performance as the result of the interaction between individual, task, and environmental characteristics [15]. In the GSTM, the individual characteristics describe factors that relate to the individual qualities of athletes, which are divided into five categories: anthropometrics, physiological, technical, tactical, and psychological characteristics. Task characteristics represent the sport and its specific demands, and environmental characteristics include the surroundings of the athlete, e.g., sociodemographic characteristics of the athlete. The interactions between the individual characteristics, the task and the environment change over time under the influence of individual maturation, learning, and training processes. It is of essence to reveal the ‘key’ characteristics related to performance and to obtain insight into these developmental processes as this helps to improve the development of young talented athletes toward expert performers. Here, a longitudinal approach is expected to be more informative than a cross-sectional approach [14, 16]. Longitudinal studies involve the repeated assessment of the same individuals over a longer time period (e.g., 6 months). Such studies have the advantage of capturing the long-term changes in athletes’ performance characteristics and relating them to future career-related outcomes to discover which characteristics could be important for successful future performance [2]. Furthermore, longitudinal data can contribute to the creation of realistic goals and training procedures in talent development [17]. For example, a study in handball (n = 94, age 13–17 years) revealed how multiple individual characteristics developed over time in different age- and performance groups [18]. This study showed the importance of specific performance-related characteristics and how the characteristics as well as their importance developed over time in this specific sport, helping coaches to evaluate players.

This review intends to gain further insight into the outcomes of not only longitudinal but also multidimensional approaches and to provide directions for future talent research by means of a focused study on individual characteristics in racket sports. Racket sports are generally considered early specialization sports (at 6–10 years of age), and peak performance is frequently reached relatively late (25–35 years of age) and can last a long period (up to 15 or 20 years) [19]. As such, players aiming for the elite level and other stakeholders connected to this developmental process (e.g., parents, trainers, coaches, clubs, and associations) need to invest great amounts of resources to increase the chances of successes. Better methods to support talent identification and development within racket sports are likely to contribute to more effective and efficient TID programs. This can be of great value for both players and other stakeholders when deciding about their pathway and investments. The assessment of individual characteristics and their relationship to performance have already been reviewed for racket sports [19,20,21]. For example, instruments focusing on intellectual and perceptual abilities and coordinative skills were able to discriminate between various performance levels. However, their predictive validity was not yet confirmed. Furthermore, there was moderate evidence that assessing mental and goal management skills could predict future performance [19]. Also, the assessments of sport-specific technical skills could discriminate different performance levels and predict future performance in TID activities in different sports [21]. Moreover, there was strong evidence that technical and tactical skills differentiate performance levels specifically in tennis [20]. Nevertheless, these reviews emphasized that the individual performance-related characteristics were mostly measured in isolation and/or on a single measurement occasion. This indicates that a limited number of previous studies used a multidimensional and/or longitudinal research design.

In the recent past, more and more multidimensional and longitudinal studies have been conducted. However, to the best of our knowledge, there is no comprehensive overview of empirical/data-driven multidimensional and/or longitudinal research in racket sports. Synthesizing the knowledge on talent development and the individual characteristics in racket sports specifically using multidimensional and/or longitudinal insights would allow for a better understanding of the developmental processes and help to identify and guide young talented players to using their full potential. Therefore, this systematic review aims to provide an overview of empirical/data-driven multidimensional and/or longitudinal research in talent development within the field of racket sports. It intends to gain further insight into the outcomes of multidimensional and longitudinal approaches for talent identification and development in racket sports and provide directions for future talent research by means of the following research question: Which (set of) individual performance-related characteristics can explain performance outcomes in young talented racket sport players? This work contributes to a better understanding of what can be learned from the current literature, identify gaps and provide possible directions for future research.

Methods

Search Strategy and Eligibility Criteria

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed when conducting and reporting this review [22]. Electronic database searches were conducted in PubMed, Scopus, SPORTDiscus, and Web of Science [23]. The search was limited to peer-reviewed papers published in English from January 2000 until the 12th of August 2022. Search terms for all databases represented the following racket sports: tennis, table tennis, badminton, squash and padel. Additional search terms represented the concepts of talent, talent identification, and talent development as well as sports performance. Because the word “squash” has multiple meanings, some terms concerning fruit, vegetables, and plants were added to the search string using the operator NOT. In summary, the search for articles contained the following terms:

(“racquet sport* (MeSH)” OR “racket sport*” OR racquetball OR “racquet ball” OR racketball OR “racket ball” OR tennis (MeSH) OR “table tennis” OR squash OR badminton OR padel).

AND.

(Talent* OR aptitude*(MeSH) OR gift* OR assess* OR endowment* OR select* OR scout* OR expert* OR elite OR excellen* OR success* OR perform* OR identif* OR develop*).

NOT.

(pumpkin* OR intervention* OR vegetable* OR fruit* OR plant*).

Articles were included if they (1) were original articles containing an empirical/data-driven study using inferential statistics, (2) focused on the identification or development of talented young players in at least one of the major racket sports (i.e., tennis, table tennis, badminton, squash, and padel), and (3) included a multidimensional and/or longitudinal approach. Multidimensionality was met when at least two of the five individual characteristics (i.e., anthropometric, physiological, technical, tactical, and psychological) of the GSTM were covered in a study to compare athletes of different performance levels (e.g., elite versus sub-elite) and/or to explain performance (e.g., rating score or ranking). With this, the ‘psychological’ category covered both the psychological and cognitive performance determining factors [24] and the technical category included both technical skills (e.g., stroke velocity and accuracy) and (perceptual-)motor skills (e.g., eye-hand coordination). Study designs covering multiple measurements during a minimum time period of 6 months were defined as longitudinal studies. In both the multidimensional and longitudinal studies, the measure(s) had to be related to (any aspect) of talent or sport-specific performance outcomes. Exclusion criteria applied in this review were articles (1) concerning factors beyond the individual (i.e., task or environmental characteristics), (2) with insufficient relationship between individual characteristics and (a measure of) talent or performance (e.g., relationship between individual characteristics), (3) including solely group comparisons regarding sex (i.e., male versus female), age (i.e., younger versus older), maturation (e.g., early versus late maturing), handedness (i.e., left-handed versus right-handed), sports (e.g., racket sports versus swimming and judo) or countries (e.g., German versus Dutch athletes) without a clear performance outcome, (4) concerning the comparison of different experimental conditions or intervention studies, and (5) focusing on non-healthy or injured participants. Furthermore, duplicates and articles without full-text access were also excluded. Titles, abstracts, and full-text articles were screened by four authors (SN, TK, JM, IF) using the web tool “Rayyan” [25]. If judgment differed between the authors, articles were discussed within the research group until consensus was reached.

Data Extraction/Synthesis

Study characteristics were manually extracted into custom Excel workbooks [26]. The dataset included the following information regarding the article: name of the authors, publication year, sport(s) investigated, sample’s country of origin, sex, size and age, study design reported, dimensions measured, measurements conducted, and general findings. Subsequently, the samples’ performance level was determined based on the method of Swann et al. (2015) [27]; samples were classified as semi-elite, competitive elite, successful elite, or world-class elite athletes.

Quality of Evidence

The methodological quality of the articles included was evaluated using a modified checklist based on the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement. Modifications were based on the adaptations by Koopmann et al. (2020) [21]. Articles were assessed based on a 16-item list including a study’s (1) title and abstract, (2) scientific background and rationale, (3) objectives and hypotheses, (4) information on data collection, (5a) participant information, (5b) participant selection, (6) outcome variables, (7) statistical methods, (8) missing data, (9) main results, (10) results reported in statistical terms, (11) sources of bias, (12) post hoc comparisons, (13) summary of key results, (14) limitations, (15) interpretation of results, and (16) generalizability. The outcome was reported by allocating “0” (does not fully meet criteria), “1” (meets criteria), or “NA” (not applicable). The methodological quality was independently assessed by two researchers (SN, WE) and discussed until consensus was reached. When there was doubt, the research group was consulted. A total score was calculated by summation of the scores on each item. Percentage scores were calculated as the final score, calculated by dividing the total score by the number of relevant scored items (i.e., NA items were not included). Articles were categorized as having a low, moderate, or high methodological quality based on the percentage ≤ 60%, 61–79%, and ≥ 80%, respectively. These thresholds are in line with the cut-off scores used by recent reviews in sports science [19, 21, 28]. Subsequently, information on the studies’ quality and their findings were combined to rate the level of evidence for the dimensions within different sports. The level of evidence was categorized as strong (i.e., three studies of high OR five studies of moderate quality with/without statistically significant effects), moderate (i.e., two studies of high OR three studies of moderate quality with/without statistically significant effects), limited (i.e., one study of high OR two studies of moderate quality with/without statistically significant effects) or conflicting (i.e., < 2:1 ratio between studies with and without statistically significant effects) [19, 21].

Results

The systematic search yielded 28,932 articles from all four databases (Fig. 1). After removing duplicates (n = 7543), excluding articles based on automatic eligibility (n = 541), title and abstract (n = 20,542), and irretrievability (n = 20), 286 articles remained for full-text screening. Subsequently, 252 articles were excluded because the study design was not multidimensional or longitudinal (n = 102), the article had an insufficient relationship to talent or high-performance (n = 69), the study design of the article was ineligible (n = 66), the article was not written in English (n = 11) or the article was outside the scope of the five individual characteristics (n = 4). Articles with only semi-elite athletes were also eliminated considering the insufficient representation of talented or high-performing athletes (n = 3). One additional article was proposed during the review process which met the inclusion criteria. Finally, 32 articles remained for inclusion and were of low (n = 1), moderate (n = 6) and high (n = 25) methodological quality. The included articles were labeled as multidimensional (n = 15) and/or longitudinal (n = 17, including 14 studies using a combined multidimensional and longitudinal approach) and comprised tennis (n = 19), table tennis (n = 7), badminton (n = 4), squash (n = 2), and padel (n = 1). The distribution of the individual characteristics taken into account within the multidimensional studies is presented in Additional file 1: Table S1 of the supplementary files. The characteristics of the multidimensional and longitudinal studies can be found in Tables 1 and 2, respectively. Table 3 presents the results of the methodological quality check for all articles. In the following sections, the studies’ findings regarding various characteristics are presented starting with the characteristic investigated the most and ending with the one studied the least.

Fig. 1
figure 1

Flow chart systematic search. *If not longitudinal

Full size image
Table 1 Multidimensional studies included in the systematic review
Full size table
Table 2 Longitudinal studies included in the systematic review
Full size table
Table 3 Results of the quality check using a modified checklist based on the STROBE statement
Full size table

Multidimensional Studies

Table 1 presents the included articles (n = 15) using a multidimensional approach. Twelve studies included measurements of two characteristics. Of these, six studies measured a combination of physiological and technical characteristics, and five studies focused on the combination of anthropometric and physiological characteristics. The combination of technical and tactical characteristics was considered within one study.

There was strong evidence that better physiological characteristics (i.e., aerobic fitness/endurance) and technical characteristics (i.e., serve and stroke velocity) are associated with or can explain higher performance in tennis [29,30,31,32,33]. These findings were confirmed in all studies using univariable [31, 33] and multivariable statistical approaches [29, 30, 32]. Limited evidence was found that better physiological characteristics (i.e., upper limb and lower limb angular velocities) are related to a higher performance level in table tennis using univariable statistics since only one study concerning this characteristic could be included for this sport [34].

Studies including anthropometrics and physiological characteristics revealed conflicting evidence for body composition and moderate evidence for physiological characteristics (i.e., aerobic capacity) as determinants for performance in tennis [35, 36]. Also, within these studies, better aerobic parameters appear to be associated with higher performance. Moreover, limited evidence was found for badminton, squash, and padel concerning anthropometrics and physiological determinants as only one article could be included for these characteristics per sport [37,38,39]. In general, a trend can be recognized that better physiological outcomes are related to a higher performance level. Such a trend was not visible regarding the anthropometric outcomes. All studies focusing on anthropometrics and physiological characteristics used univariable statistical approaches.

Finally for the studies investigating two characteristics, the combination of technical and tactical characteristics was evaluated in one study. As such, the level of evidence for these characteristics is limited. This study showed that higher-level players outperform lower-level players in technical and tactical skills in squash [40]. The findings in this study were based on a univariable statistical approach.

Three studies included measurements of three performance characteristics. Of these, two studies in tennis measured a combination of anthropometric, physiological, and technical characteristics [41, 42]. Again, a trend was found that performance is associated with anthropometric, physiological, and technical characteristics. Higher-level players tend to outscore lower-level players on various anthropometric, physiological, and technical outcomes. Based on the results of these two studies, the level of evidence must be classified as limited for these findings. Both studies included univariable statistics but one study only found a trend and no statistically significant differences between performance groups [41].

Another three-dimensional study was conducted in badminton including anthropometric, physiological, and psychological characteristics [43]. This study used a multivariable statistical approach to predict players’ performance level (i.e., elite, sub-elite, or novice). The combination of anthropometric, physiological, and psychological data classified 100% of the players correctly while 80% were cross-validated correctly. Since this was the only study in badminton evaluating these characteristics, the level of evidence is limited.

Longitudinal Studies

Table 2 presents the included articles (n = 17) using a longitudinal approach. Three articles followed longitudinal and unidimensional approaches with one study assessing physiological characteristics in tennis [44], one study analyzing psychological characteristics in table tennis [45], and one study evaluating anthropometric characteristics in tennis [46]. All studies found statistically significant effects for the respective characteristics. That is, higher values for maximum oxygen uptake (VO2-max) in a cross-correlation analysis [44] and body mass index (BMI; assumed to be reflecting more muscle mass) in a polynomial regression model [46] are related to more success at the elite level in tennis. In table tennis, players on the international and national level showed higher values for motivation, coping skills, and stress tolerance compared to regional level players in an analysis of covariance (ANCOVA) [45]. Since there was only one study per characteristic, the level of evidence for all three of them is classified as limited [44,45,46].

Nine longitudinal studies included measurements of two characteristics. Most studies measured a combination of anthropometric and physiological characteristics (n = 6), of which five were conducted in tennis while one included badminton players. There was strong evidence that anthropometric (i.e., cross-sectional area of the quadriceps femoris muscle) and even more so physiological characteristics (i.e., lower body power [e.g., sprinting, jumping] and upper body power [e.g., medicine ball throwing]) were related to success and performance level in tennis [47,48,49,50,51]. However, it must be noted that there were some conflicting findings when evaluating characteristics in detail. For example, one study found relationships or differences for sprinting or jumping tests [48] while another did not [49]. As there was only one longitudinal study presenting results for physiological characteristics in badminton (i.e., jumping, badminton-specific speed, and endurance), this evidence was classified as limited [52]. All six studies used multilevel or multivariate analyses.

Two studies investigated a combination of physiological and technical characteristics. Both studies investigated whether perceptual-motor skills could predict future performance in competitive elite youth table tennis players using multilevel or multivariate analyses [53, 54]. While the assessments for technical characteristics (e.g., throwing a ball, aiming at target) showed statistically significant effects in both studies and thus represent moderate evidence, physiological characteristics (i.e., sprint, agility, jumping) showed effects only in one study and the evidence must be classified as conflicting accordingly. As above, it must be noted that there were more conflicting findings once characteristics were analyzed in deeper detail with respect to specific assessment and measurement methods. For example, one study found a positive effect for eye-hand coordination [53] while another did not [55]. Similarly, two studies found positive effects for sprinting in table tennis [54, 55] while two other studies did not find such effects [53, 56].

Finally for the longitudinal studies examining two characteristics, one study investigated a combination of anthropometric and technical characteristics in tennis using a multivariate analysis and found that future elite players were significantly taller and heavier compared to future competitive players. Also, ball speed and accuracy were significant predictors of current and future performance [57]. As this was only one study, the evidence was classified as limited.

Four longitudinal studies included measurements of three characteristics. Of these, three studies, two in tennis and one in table tennis, measured a combination of anthropometric, physiological, and technical characteristics [56, 58, 59] while one study combined anthropometric, physiological, and psychological characteristics in table tennis and badminton [60]. The first three studies found statistically significant effects for physiological (i.e., sprinting, jumping), and technical characteristics (e.g., ball throw) using a univariate as well as a multilevel analysis [56, 58, 59]. Conflicting findings were found for the predictive value of anthropometrics (i.e., height, weight) [56, 58, 59]. Accordingly, overall, there was moderate evidence for tennis and table tennis combined and limited evidence for each sport separately. In this context, it must be noted that anthropometric assessments were often used not as primary variables but rather as additional variables (i.e., to assess maturation) for the interpretation of, e.g., physiological or technical characteristics [36, 42, 43, 48,49,50,51, 55, 57, 58]. The last study with three characteristics evaluated the longitudinal development of anthropometric, physiological, and psychological characteristics over a 2-year period in male table tennis and badminton players using a discriminant analysis and multilayer perceptron neural network. The study found that all players improved in all aspects [60]. Accordingly, there was limited evidence that players get taller, heavier, fitter, and technically better over time.

Only one multidimensional and longitudinal study investigated four performance-related characteristics using correlation analysis. This study explored whether multidimensional profiling could be useful in predicting table tennis player’s current and future performance level one year later [55]. Statistically significant Spearman rank-order correlations were found for physiological (i.e., sprint) and a few psychological characteristics (e.g., work engagement, self-regulation). As there was only one study, the evidence level was classified as limited.

Discussion

This systematic review aimed to provide an overview of empirical/data-driven multidimensional and/or longitudinal research in talent development within the field of racket sports. It intended to gain further insight into the outcomes of multidimensional and longitudinal approaches for talent identification and development in racket sports and provide directions for future talent research.

The findings show that multidimensional and longitudinal studies are being conducted in racket sports, especially in tennis and table tennis. However, despite the relatively high number of multidimensional and longitudinal studies (n = 32), it remains difficult to draw strong conclusions. This is mainly due to the lack of uniformity in various aspects of the studies. There are differences in the (1) operationalization of constructs (i.e., variables measured), (2) measurement instruments (i.e., how to measure these variables), (3) study designs, and (4) statistical approaches. Furthermore, despite the mostly moderate to high ratings for methodological quality, the study designs and statistical approaches that have been applied did not always seem to maximize the datasets’ added value for talent research. In approximately 50% (14/29) of the included multidimensional studies, a univariate analysis was chosen, where a multivariate analysis perhaps could have been more informative to further unravel the complexity of talent and its multidimensional nature. Although the chosen analyses were appropriate for the aims of the individual studies, optimally, statistical analyses should utilize the potential of multidimensional data and conduct multivariate analyses in talent research that comprise a set of performance characteristics. Consequently, characteristics are not only investigated in isolation but specifically in combination to reveal potential interaction effects. Such an approach offers the possibility to shed light on the so-called ‘compensation phenomenon’ [3, 61]. When players score poorly on certain performance characteristics, they can potentially compensate for this by well-developed other characteristics. This kind of information helps to better understand talent identification as well as developmental processes. In addition, statistical analyses should utilize the potential of longitudinal data and conduct not just cross-sectional analyses, but analyses that capture the performance characteristics on several points in time. Here, a focus on the difference in scores between those time points also appears valuable. That way, essential information can be revealed about improvement, stability or even decrement of a player’s performance characteristics in certain periods during their sports career [12].

Future research should aim to find best-practice assessments for various performance characteristics and use adequate (multivariate) statistical analyses to carefully interpret the results in detailed ways. A recent example for handling multidimensional data can be found in research reported by Robertson et al. (2022) who followed these steps in a best-practice manner [43]. The authors analyzed various variables regarding three different characteristics in three different sub-samples using a multivariate analysis of covariance (MANCOVA) and a discriminant analysis. Examples of best practice in longitudinal research can be found in studies applying multilevel modeling (see for an overview Elferink-Gemser et al. 2018 [4]). Multilevel modeling is an extension of multiple regression, which is appropriate for analyzing hierarchically structured data [62]. A two (or more) level hierarchy can be defined, with the repeated measurements (i.e., level-1) nested within the individual players (i.e., level-2). An advantage of using a multilevel regression modeling approach is that both the number of measurements and the temporal spacing of the measurements may vary between players [63]. This addresses one of the main challenges in longitudinal research, i.e., how to deal with incomplete datasets which are quite common in long-lasting research with humans. A multilevel model not only describes underlying population trends in a response (the fixed part of the model), but also models the variation around this mean response using the time of measurement and individual differences (the random part). In addition, researchers may need to use statistical analyses focusing on individual differences and development instead of group comparisons as world-class athletes are by definition outliers within statistical analyses.

While acknowledging the limitations in bringing together the multitude of studies in the current review, several relevant trends can still be observed. It became clear that almost all studies measured physiological characteristics in combination with either anthropometric, technical, or psychological characteristics. Anthropometric characteristics are frequently used for the interpretation of other outcomes. For example, an athlete’s maturation status (e.g., age at peak height velocity, APHV) can be determined based on anthropometric measures [64, 65]. Consequently, these measures provide information for the interpretation of other characteristics [48, 49, 51, 55, 57]. To illustrate, strong evidence was found that body height is related to serve speed in tennis [41, 66,67,68,69], while other studies found that serve speed was related to overall performance in tennis [32, 33, 42]. Together, this example shows how anthropometrics may be indirectly related to performance.

In some instances, the studies allowed for an insight into the individual racket sports in line with their task-specific demands. In tennis, strong evidence was found for physiological characteristics in both multidimensional [29,30,31,32,33, 35, 36, 41, 42] and longitudinal studies [44, 49, 59]. Physiological characteristics related to motor coordination, sprint, strength, and endurance were reported to be beneficial for progressing through TID programs [58] and to increase a player’s chances of achieving expert performance. In badminton, more advanced players performed better on motor fitness [37], explosive power, flexibility, endurance, and speed [43]. When investigated longitudinally, speed and endurance improved with age, and youth players (U19) reached comparable speed levels to those of world-class players [52]. In table tennis, more advanced players performed higher joint torques at higher racket speeds [34]. Also, current performance and performance progression were found to be related to sprint speed in youth table tennis players [55]. For squash and padel, limited evidence was found for the relationship between physiological characteristics and performance given that only two studies were conducted [38, 39].

Regarding technical characteristics, limited evidence was found for the different types of multidimensional and longitudinal studies separately, while moderate evidence was found when all were combined. Moderate to high correlations were found between technical characteristics and performance measures [29, 32], and technical skills were able to discriminate between performance levels in tennis [29]. This is in accordance with a previous systematic review of research on technical and tactical skills in tennis which found technical skills (e.g., ball velocity and ball accuracy) to be discriminative [20]. Also, differences in technical characteristics were found between performance levels in squash [40]. In both tennis and table tennis, technical characteristics were able to predict future performance [20, 53, 54, 56, 59] and were considered important for progression through several stages of a TID program [58]. Similarly, this was shown for, among others, perceptual abilities and coordinative skills in a previous systematic review [19]. In most sports, technical and tactical characteristics are very strongly connected. Technique often plays a functional role in executing a tactical decision to reach a certain goal [20]. However, only a single study investigated tactical characteristics using a multidimensional and/or longitudinal approach. Therefore, only limited evidence was found for the relationship between tactical characteristics and performance [40]. It is important to note that a previous review found many cross-sectional and unidimensional studies investigating (technical and) tactical or perceptual-cognitive skills in racket sports [20]. This fact further emphasizes the need for longitudinal research.

Although only limited evidence was found for psychological characteristics in the different types of multidimensional and longitudinal studies separately, it must be mentioned that several articles did find that psychological characteristics were related to (future) performance [43, 45, 55]. These studies used different questionnaires (e.g., Psychological Characteristics of Developing Excellence Questionnaire, version 2 [PCDEQ2], or Sport Motivation Scale [SMS]) to assess various psychological concepts and skills. For example, this included intrinsic and extrinsic motivation, self-regulation, coping skills, and mental toughness. These findings are similar to a previous systematic review, finding moderate evidence for the assessment of mental and goal management skills in racket sport players [19].

Some limitations of this systematic review must be acknowledged. First of all, this review focused only on one part of the GSTM [13] as it isolated the athlete’s individual characteristics and did not include valuable information beyond the individual (i.e., task and environmental characteristics). For example, sociodemographic characteristics, such as family support or coaching situation could have an influence on whether the athlete has the possibility to play at an elite level [6, 13, 70,71,72]. The inclusion of this information is recommendable for a better understanding of the complexity of talent identification and athlete development. Secondly, some measures were difficult to categorize into a single individual characteristic because of the complex interaction between characteristics. To illustrate, eye-hand reaction time consists of both a psychological/cognitive and a technical-motor component [73] and could therefore be categorized as a technical or psychological characteristic. Depending on this categorization decision, a study could have been considered unidimensional instead of multidimensional and consequently been excluded for this reason. Thus, it is important for all researchers in the field to create and report their studies using generic terminology and categorization so that all relevant information from the literature can be captured and combined. This must include specifically the study’s design, statistical approach, and participants’ age and performance level. Thirdly, the findings of this review may have been influenced by a publication bias. Although the most commonly used databases for sport settings were used [23], a wider search among more databases including not only English-language studies and other study designs and/or grey literature might yield new insights. For example, no studies on coaches’ perspectives were included, even though they can be of added value when identifying important individual characteristics and their relationship to performance [74,75,76]. Moreover, publication bias within the empirical/data-driven studies might be due to favoring statistically significant and positive results over null or negative/conflicting results. Particularly given the comparison and combination of results of different studies and seeing partially conflicting findings (e.g., for sprinting), this may be relevant. Thus, all researchers should strive to improve publication procedures using new and more transparent approaches (e.g., pre-registration). Fourthly, although racket sports share certain task similarities, there are still differences in demands between them, e.g., physiologically [76, 77], and the settings in which the measurements take place (i.e., a laboratory versus ecological-valid context, e.g., in a competitive setting) must be considered when interpreting and transferring results. Consequently, generalization of findings from one racket sport to another, as well as from certain (assessment) settings to others, must be done with caution.

Conclusions

In conclusion, this systematic review provides an overview of talent research using multidimensional and/or longitudinal approaches within racket sports. Despite the apparent challenges of bringing together the variety of current multidimensional and longitudinal studies, this review revealed pieces of relevant information on which individual performance-related characteristics could help explain performance outcomes in young talented racket sport players. Depending on the specific sport of interest, the current literature provides both practitioners and researchers some guidance on what characteristics to include in their decision-making processes in TID contexts. Future research should conduct more multidimensional and longitudinal studies combining various individual and also environmental characteristics. Characteristics should be assessed using (standardized) best-practice methods to allow for better comparisons and combinations of studies. Also, data should be analyzed using adequate (multivariate) analyses to effectively take advantage of multidimensional and longitudinal approaches’ added value. When presenting their findings, researchers should use generic terminology and extensively describe the study’s methodological approach and sample. All in all, this helps to continue unraveling the concepts and processes underlying successful talent identification and development in (racket) sports.