Introduction

Within many youth sports contexts, athletes are often organised into annual age groups using specific cut-off dates (e.g. 31st of December) to ensure equal and fair competition levels for young athletes (Cobley, Baker, Wattie, & McKenna, 2009; Helsen, Winckel, & Williams, 2005; Wattie, Cobley, & Baker, 2008). However, the annual age-grou** still allows for chronological age differences of up to 12 months, leading to chronological age advantages (Cobley et al., 2009). While the birth distribution is evenly split in the quarters of the year, unbalanced distributions can be observed in various sports contexts with an overrepresentation of relatively older athletes whose birth months are closer to the cut-off date for the competition categories within the selection year (Helsen et al., 2005; Musch & Grodin, 2001). This is described and well-documented as relative age effects (RAE) and can lead to misjudgements in the identification and selection of talents (Cobley et al., 2009; Difernand et al., 2023; Gil, Bidaurrazaga-Letona, Larruskain, Esain, & Irazusta, 2021).

Since the 1980s, many studies identified the prevalence of RAE in various physically demanding sports (Cobley et al., 2009; Lames, Augste, Dreckmann, Görsdorf, & Schimanski, 2008). In swimming, the RAE is prevalent among age-group athletes in several countries (Australia: Abbott et al., 2020; Cobley et al., 2018, 2019; Germany: Altmann & Sperling, 2013; Staub, Stallman, & Vogt, 2020; Thiel & Spitzpfeil, 2014; Portugal: Costa, Marques, Louro, Ferreira, & Marinho, 2013). In the recent review of Lorenzo-Calvo et al. (2021), the researchers observed the high prevalence of the RAE in up to 60% of the cases analysed in age-group swimming. The magnitude is more accentuated in male and younger swimmers but decreases as the chronological age of the swimmers increases. The impact of the RAE on competitive performance is related to the demand for strength in the event, as the performance in simultaneous strokes, in shorter events, and on swimmers in the post-adolescence period. The authors conclude that the RAE in swimming relies on the individual (sex), environmental (starting field in a certain age-group), and task constraints (competitive event; Lorenzo-Calvo et al., 2021).

Based on the results presented above, the prevalence of the RAE indicates that the talent identification (TID) and development systems in youth swimming are biased against relatively younger athletes. All findings show that consensually this means having a lower likelihood of being selected and thus having access to a higher level of coaching, training, and other talent-promoting factors supporting the development to an elite level.

Young athletes also have an advantage in TID when their biological maturation is more developed, primarily in male athletes (Deprez et al., 2013). While annual age-grou** permits chronological age advantages up to 12 months, biological age differences are potentially up to five years during rapid maturation, where timing and tempo vary between individuals (Malina, Bouchard, & Bar-Or, 2004). Furthermore, the maturation of different systems in the body proceeds independently of chronological age (Malina, Rogol, Cumming, Coelho-e-Silva, & Figueiredo, 2015). Therefore, the chronological age represents a weak indicator of biological maturity (Beunen, Rogol, & Malina, 2006), which is why the relation between the RAE and the biological maturity status should be considered. A lower biological maturity status is, mathematically, an explanation for the influence of the RAE on TID in sports (Romann & Cobley, 2015), which refers to the ‘maturation-selection hypothesis’ (Baker, Janning, Wong, Cobley, & Schorer, 2014). The assumption is that relatively older athletes have an increased likelihood of advanced normative anthropometric development compared with relatively younger counterparts (Malina et al., 2015). The development of advanced anthropometric characteristics, e.g., greater body height (BH) and lean body mass (BM), are predictive of better physical characteristics such as muscular strength, power, speed, and endurance (Baker et al., 2014; Cobley et al., 2009; Malina, Chamorro, Serratosa, & Morate, 2007; Viru et al., 1999). The scientific results that relatively older athletes are taller and heavier compared to their relatively younger counterparts and the influence of anthropometric characteristics on the RAE were demonstrated in youth soccer and youth alpine skiing (Deprez et al., 2013; Gil et al., 2021; Müller, Gehmaier, Gonaus, Raschner, & Müller, 2018; Müller, Müller, Hildebrandt, Kornexl, & Raschner, 2015b). The interaction of immediate performance advantages due to such anthropometric characteristics has also been shown in analyses of swimming performance (Aspenes & Karlsen, 2012; Zacca et al., 2020).

In recent years, there has been an increasing research interest in RAE concerning the biological maturity status (Müller et al., 2015b; Skorski, Skorski, Faude, Hammes, & Meyer, 2016). To estimate the biological maturity status, the indicator age at peak height velocity (APHV) can be used, which represents the time of maximum growth during puberty (Mirwald, Baxter-Jones, Bailey, & Beunen, 2002; Müller, Müller, Hildebrandt, Kapelari, & Raschner, 2015a). In youth soccer (Deprez et al., 2013) and youth alpine skiing (Müller et al., 2015b), researchers measured APHV. They reported no statistical difference regarding the biological maturity status in athletes born within one year. Some researchers concluded that relatively older athletes had an increased likelihood of selection independent of their biological maturity status. In contrast, relatively younger athletes often only had a chance of being selected if they had a similar biological maturity status to their older counterparts (Gil et al., 2021, Skorski et al., 2016).

To summarise, it is well-known that relatively younger and less mature swimmers suffer performance and selection disadvantages for representative competitions (Cobley et al., 2018). Such disadvantages may persist until growth deceleration (post-peak height velocity) and maturation is reached (Baxter-Jones et al., 2020). It can be assumed that there is a ‘waste of potential’ of genuinely skilled athletes due to the prevalence of the RAE, which represents a systematic error in the TID (Jiménez & Pain, 2008; Lames et al., 2008).

In order to shed more light on the relationship between RAE and maturation in the talent identification process, this study examines a cohort of young swimmers participating in an annual event for talent identification. According to current knowledge, such a study has yet to be published in youth swimming. The objectives of the present study were to determine (1) the prevalence and the magnitude of the RAE and (2) its relation to anthropometric characteristics and the biological maturity status, according to sex, by German youth swimmers aged 9 to 13. We hypothesise that (1) there is no equal distribution of birth quarters in the sample and that (2) the measures of physical development and maturation collected (BH, BM, and APHV) are unrelated to birth quarters.

Materials and methods

Samples and data preparation

The study was conducted following the guidelines of the Declaration of Helsinki as well as following an institutional ethical approval (German Sport University Cologne; Nr. 045/2018). To determine the prevalence and magnitude of and its connection to the biological maturity status in (German) youth swimming, the results of a standardised motoric test (‘Landesvielseitigkeitstest’, LVT) of swimmers born in the years from 2005 to 2009 between 2017 and 2018 were used. The LVT is a predetermined competition for talent identification (TID) on the federal level, which is implemented on the state level. The qualification for participating at the LVT is based on multiple competition performances. Cross-sectional data from the two following years provided by the Swimming Association of North Rhine-Westphalia (‘Schwimmverband Nordrhein-Westfalen e.V.’, SV NRW) were used to increase the number of athletes and set a representative sample of subjects. The data were systematically screened for doubles. Multiple cases of persons with identical names were identified and marked as different (Staub et al., 2020). Incomplete and illogical data were not considered. In total, 650 swimmers remained for analyses: 273 male swimmers (chronological age: 11.02 ± 1.30 years; range: 8.6–13.4 years) and 377 female swimmers (chronological age: 11.00 ± 1.25 years; range: 8.5–13.4 years). The provided data that was used for analysis and included date of measurement, swimmer’s date of birth, sex, body mass (BM), body height (BH), and sitting height (SH).

Procedures

The cut-off date for age-grou** in German youth swimming is set for January 1 to December 31 (Staub et al., 2020); thus, the birth months of the swimmers were split into the four age quartiles as follows: January–March were categorised as age quartile 1 (Q1), April–June as quartile 2 (Q2), July–September as quartile 3 (Q3) and October–December as quartile 4 (Q4).

To judge the prevalence and magnitude of the relative age effect (RAE), the distributions of age quartiles were compared to actual distributions of birth in the German population from 2005–2009. These birth data were accessed from the German Bureau of Statistics (DESTATIS, 2021). Across the mentioned years of birth of the sample, 3,391,021 live births occurred and were evenly distributed (Q1: 24.13%, Q2: 24.71%, Q3: 27.10%, Q4: 24.07%).

The biological maturity status was assessed by the non-invasive method of calculating the age at peak height velocity (APHV) proposed by Mirwald et al. (2002). The sex-specific prediction equations include anthropometric parameters like body mass (BM), body height (BH), and sitting height (SH) that were obtained from the data provided. The procedure of measurement of the anthropometric parameters follows the test manual of Altmann (2017). The calculations of leg length as the difference between BH and SH and the actual chronological age at the time of measurement were included in the equations. Based on this, the maturity offset, meaning the time before or after individual peak height velocity, could be assessed to calculate the predicted APHV as the difference between chronological age and maturity offset (Mirwald et al., 2002).

Statistical analysis

Descriptive data calculated for the samples included frequency distribution, relative frequencies (%), mean value and standard deviation (M ± SD). Concerning the prevalence and magnitude of the RAE, χ2 were used to determine differences between the observed and normatively expected distributions. Cramér’s V identified the magnitude of effect size between Q1 and Q4 frequency counts. Magnitude estimates ranging between 0.06 < V < 0.17 indicated a small effect size, 0.17 < V < 0.29 a medium effect, and V ≥ 0.29 a large effect size (Cramér, 1999). The distribution of birth quarters was compared to examine potential patterns or variations. The analysis focused on assessing whether there were significant differences in the distribution of births across different quarters of the year. Odds ratios (OR) were computed (Q1 vs Q4; Q2 vs. Q4; Q3 vs Q4) according to sex using logistic regression to quantify these differences, allowing us to determine the odds of being born in a specific quarter compared to others as proposed by Cobley et al. (2009). Additionally, 95% confidence intervals (CI) were calculated to provide a range of values within which the true effect size is likely to fall. The effect size was further assessed using Cohen’s d, with categorisation into small (0.2 ≤ d < 0.5), medium (0.5 ≤ d < 0.8), and large (d ≥ 0.8) effect sizes. This comprehensive approach aimed to elucidate the statistical significance and practical relevance of any observed differences in the distribution of birth quarters.

The normal distribution of the anthropometric and maturity-related parameters was tested using a Q–Q plot, separated by sex (Upton & Cook, 2014). To assess the differences in the BM, BH, and APHV between the four age quartiles and separated by sex, one-way analysis of variance (ANOVA) was conducted (dependent variables, BM, BH, APHV; independent variable: age quartile). The variance homogeneity was tested using the Levene test. Bonferroni corrected pairwise comparisons were used for post hoc analysis (Upton & Cook, 2014). The significance level was set at α = .05, except for the Levene test (α = .1). All calculations were performed using Microsoft Excel Version 2021 (Microsoft Corporation, Redmond, WA, USA), IBM SPSS Statistics Version 23 (IBM, Armonk, NY, USA), and RStudio Version 1.2.5001 (RStudio Inc, Boston, MA, USA).

Results

Table 1 summarises the age quartile distributions, χ2 test statistics, effect size estimations, categorisation, and OR analyses according to sex. Findings identified that regardless of sex, relative age effects (RAE) were prevalent with medium effect sizes. The results highlighted significant variations in the distribution across age quartiles for male and female swimmers, as evidenced by the χ2 test. For males, the χ2 value was 28.18 (p < .001), while for females, it was 47.64 (p < .001), indicating highly significant differences. Further exploration through OR analysis revealed noteworthy findings. Effect size measures, including Cramer’s V (0.18 for males, 0.20 for females) and Cohen’s d, indicated practical significance, with small to medium effect sizes observed. These findings offer valuable insights into the complex interplay of age and gender among swimmers. As an example, in Fig. 1 we show the distribution of the participants included in our sample in regard to 100 m butterfly (male) and 100 m individual medley (female) in the national top-100 rankings according to quartile.

Table 1 Distribution, Chi-square and odds ratio analysis of the swimmers (male, 11.02 ± 1.30 years of age; female, 11.00 ± 1.25 years) according to age quartile and gender
Full size table

Table 2 presents anthropometrical and maturity-related characteristics across the age quartiles for both sexes (APHV: male 13.31 ± 0.61; female 11.56 ± 0.46). The one-way ANOVA has not revealed significant differences regarding BH, BM, and APHV between the relative age quarters for both sexes, except for body mass in female swimmers. Female athletes born in Q1 (42.89 ± 8.41 kg) were significantly heavier than those born in Q4 (39.09 ± 8.59 kg; p = .037).

Table 2 Anthropometric data (BH, BM) and biological maturity status (APHV) of the swimmers (male, 11.02 ± 1.30 years of age; female, 11.00 ± 1.25 years) according to age quartile and gender
Full size table

Discussion

The present study aimed to (1) examine the prevalence and the magnitude of relative age effects (RAE) and (2) its relation to anthropometric characteristics and, therefore, the biological maturity status according to sex. Confirming the hypothesis, the findings revealed a relative age effect (RAE) with medium effect sizes in German youth swimming for male and female athletes. At the same time, no statistical differences were observed in body height (BH), body mass (BM), and age of peak height velocity (APHV) between the relative age quartiles for both sexes, except for BM in female swimmers. Regarding the RAE, an overrepresentation of athletes born in the first age quartile and an underrepresentation of athletes born in the last age quartile were found. This indicates that relatively older male swimmers born close to the cut-off date are almost four times more likely and females are more than three times more likely to participate in competitions for talent identification (TID) at the state level than their younger counterparts. Similar results were observed in another sample of swimmers participating in the same competition in another federal state of Germany (Saxony; n = 270; Altmann & Sperling, 2013). Furthermore, the study’s findings confirm previous investigations in youth swimming in various countries. Staub et al. (2020) investigated a German cohort of top-100 ranked swimmers between 11 and 18 (n = 3630) and revealed a significant RAE over 13 events for female swimmers up to the age of 13–15 and males up to the age of 16–18. Cobley et al. (2018) examined a cohort of participants in the Australian Age Swimming Championships aged between 12 and 18 years (n = 6014). An RAE was identified in female swimmers 12–14 and male swimmers 12–15, irrespective of stroke and distance—moreover, Costa et al. (2013) investigated a Portuguese cohort of top-50 ranked athletes between 12 and 18 years of age (n = 7813). The researchers identified a disproportionately high distribution of relatively older swimmers at 12 years for females and 12–15 years for males. In contrast to all these studies, younger swimmers were investigated in the present study. This is in line with the findings concerning the age groups 9–13 in various physically demanding sports (e.g., swimming, alpine skiing, soccer; Marapen & Low, 2015; Müller et al., 2018; Müller, Müller, Hildebrandt, & Raschner, 2016). The likelihood for a young swimmer of the first age quartile was 2.86 (male) or 2.49 (female) times higher than for a counterpart of the last age quartile. Likewise, Staub et al. (2020) and Cobley et al. (2018) identified both sexes’ similar magnitudes of RAE with large to medium effect sizes in German and Australian swimmers aged between 11 and 13 years. The results confirm that RAE is already prevalent at a younger age. Its magnitude decreases in the older age groups in all three studies, regardless of sex. Therefore, the TID system in an individual sport like swimming is biased and discriminates against relatively younger athletes. To better understand the underlying factors of selection bias, the role of the biological maturity status as a possible influential factor in TID was evaluated in the present study. The anthropometric characteristics showed no difference within the male and female samples for all parameters (BH, BM, APHV) except BM in female swimmers. Thus, maturity status was similar within the age groups. These findings indicate that the youth swimmers in this cohort are in the same maturation stage, independent of the age quartile they were born in. Consistent with the interpretive approach of Deprez et al. (2013), it seems for male and female athletes that relatively younger swimmers can counteract their relative age disadvantage regarding the RAE if they have the same anthropometry and biological maturity status as their older counterparts. Linked to this, it can be assumed that relatively younger swimmers have to be more mature than their older counterparts to participate in competitions for talent identification.

In the physical development of youth athletes, swimmers with maturity-related advantages are more likely to show better physical performances than less matured athletes as a result of maturity-related adaptations like increased androgen concentrations, fibre-type differentiation, resting adenosine triphosphate, creatine phosphate levels, and architectural development of musculotendon units (Lloyd & Oliver, 2012; Myer et al., 2011; Peña-González, Fernández-Fernández, Moya-Ramón, & Cervelló, 2018). Since swimming requires high levels of physical characteristics, especially in short-distance events, there are high demands on strength, power, and BH (Jürimäe et al., 2007; Strzala & Tyka, 2009; West, Owen, Cunningham, Cook, & Kilduff, 2011), it is a kind of sport in which the prevalence of RAE is likely to occur (Lames et al., 2008) and in which taller athletes attain advantages (Majumder & Choudhury, 2014). Based on the findings and interpretative approaches of the present study, it can be assumed that the anthropometry and the biological maturity influence the RAE and, thus, the TID in youth swimming. Further, a significantly lower BM in female swimmers born in Q4 than their counterparts born in Q1 is not necessarily a disadvantage. An optimal BM depends on the distance and stroke of swimming (Kjendlie & Stallman, 2011). In contrast, a higher BM can be advantageous if associated with greater muscle mass (Cossio Bolaños et al., 2019). Greater muscle mass increases strength and power, positively affecting sprint performance in swimming (Dopsaj et al., 2020; Garrido et al., 2010). The present study’s anthropometry and biological maturity findings align with various studies in other physically demanding sports. Deprez et al. (2013), Skorski et al. (2016) and Patel, Nevill, Cloak, Smith, and Wyon (2019) identified no significant differences in BH, BM, and APHV between the four relative age quartiles in Belgian, German, and English male youth soccer players. Further, Müller and colleagues (Müller, Gonaus, Perner, Müller, & Raschner, 2017; Müller et al., 2015b, 2016), as well as Gil et al. (2021) found no significant differences in biological maturity status between the four age quartiles in Austrian youth alpine ski racers and Spanish youth soccer players aged between 9 and 10 years. In contrast, the athletes born in Q1 were significantly taller and heavier than their relatively younger counterparts (Müller et al., 2015b).

Relatively younger soccer players were significantly more mature than their relatively older counterparts. Contrary to the findings of the present study, Müller and colleagues (2018, 2017) identified significant differences in APHV between the four age quartiles in their cohort of youth soccer players aged between 10 and 12 years of age (n = 423) or somewhat under 9 years of age (n = 222). In contrast to the interpretative approach of the present study, these results suggest that relatively older soccer players have an increased likelihood of selection independent of their biological maturation (Müller et al., 2018, 2017). A possible explanation for the difference regarding the present study is that the qualification for competitions at the regional and state level in swimming is based on (multiple) results in individual competitions (Verhölsdonk & Lennhoff, 2022). The findings of this study suggest that relatively younger swimmers who are less mature cannot compete in the standardised motoric test, which is a predetermined competition for TID. Therefore, they are denied the opportunity to participate in competitions on a higher level because they must contend with selection disadvantages. Disregarding the progression of their performance development due to a later maturation, these swimmers have fewer chances of being selected and of being identified as talented (Abbott et al., 2021).

Consequently, the results of the present study confirm that the TID system in swimming is biased. TID systems should aim to identify athletes with great potential in the long term through effective, consistent, and continuous support for success at the elite level (Rudolph et al., 2015). In this context, the RAE represents an error in the talent selection process of young athletes (Wattie & Baker, 2020). In the lower levels of TID in youth swimming, relatively younger and/or less mature athletes have fewer opportunities to be selected, getting access to a higher level of coaching, training, and other talent-promoting factors and reaching the elite level despite their talent and effort. Consequently, they often drop out of sports early and unnoticed, which can be considered a waste of the potential of genuinely skilled athletes (Jiménez & Pain, 2008).

Some recommendations are to increase greater equality of opportunity in TID for athletes and reach the full potential of relatively younger and late mature athletes (Abbott et al., 2021; Romann & Cobley, 2015). In this context, anthropometric and maturity-related characteristics should be considered more extensively in TID in youth swimming (Cobley, Romann, Javet, Abbott, & Lovell, 2020). Initially, raising the awareness of coaches, scouts, and other practitioners about the RAE and its relation to the anthropometry and biological maturation status is recommended by Cobley et al. (2009). Furthermore, tracking growth and maturation (e.g., APHV) in talent development is beneficial, and other criteria for talents (qualitative) should be considered in TID, e.g., the technical skill level (Cobley et al., 2020; Malina et al., 2017).

In team sports like soccer, bio-banding can be a possible approach to compensate for the impact of interindividual maturity-related differences, which has yet to be systematically applied (Malina et al., 2019). This strategy promotes late and early maturing youth athletes by opening the age groups according to the actual maturation status (Malina et al., 2019). A few studies have qualitatively examined the individual effects of bio-banding for soccer players from their perspectives (Bradley et al., 2019; Cumming et al., 2018) and the perspective of professional football academy staff and parents/carers (Reeves, Enright, Dowling, & Roberts, 2018). The results were consistent: early maturing players saw more significant physical and technical challenges and thus new opportunities and challenges. Later maturing players saw fewer physical and technical challenges but more opportunities to demonstrate technical and tactical skills. The corresponding results were also found in a quantitative analysis of technical–tactical key performance indicators in youth elite soccer (Lüdin, Donath, Cobley, & Romann, 2022). Applied to swimming, the bio-banding approach could create similar benefits. For example, early maturing swimmers may find new challenges in competing with older swimmers and progress beyond their physical advantage. In addition, later maturing swimmers would have a chance to compete at a higher level apart from their biological disadvantage. On the other hand, methodological limitations in determining maturity status with the APHV analysis still complicate the organisation of bio-banding in sports (Kozieł & Malina, 2018).

Recently, the Australian research group of Abbott and colleagues (2021) introduced another promising approach, the “Maturity-based Corrective Adjustment Procedures” (Mat-CAPs). It is a reasonable approach to reduce the inequality caused by RAE and maturity-related developmental differences using a mathematical formula to correct competition results in youth swimming in this respect.

To conclude, findings of our study contribute to the understanding of the RAE phenomenon, particularly in young female athletes. It should be considered in an in-depth analysis to categorise the swimmers into early, normative, and late developed. This can be used to check if the sample tends to include early developers. It is to be noted as a limitation that the association has provided the analysed data of this study. The data were collected according to a standardised procedure, but the research group could not attend to the measurement. Furthermore, for more in-depth analyses, looking at a larger sample size to calculate how the relationship behaves in individual age groups and how relative age and maturation advantages affect swimming performance would be desirable.

Our investigation provides data contributing to the research on the RAE and its relation to anthropometry and biological maturity in youth swimming. Furthermore, the RAE is prevalent in German youth swimming with the same magnitude in males and females. No significant differences exist in anthropometric and maturity-related characteristics between the four age quartiles, except the BM in female swimmers. Consequently, relatively younger swimmers may be able to counteract their relative age disadvantage if they have the same anthropometry and biological maturity status as their older counterparts.

Fig. 1
figure 1

Distribution of swimmers listed in national top-100 rankings in 100 m butterfly (male) and 200 m individual medley (female) between 2004 and 2013 according to quartile

Full size image