The Journal of Sport and Exercise Science, Vol. 8, Issue 1, 1-6 (2024)
www.jses.net
Sport psychology consultation: The attitudes of New Zealand athletes
Sam Richardson1* , Warrick Wood1, Toby Mundel1,2
1School of Sport Exercise and Nutrition, Massey University, New Zealand
2Department of Kinesiology, Brock University, Canada
Received: 13.04.2023
Accepted: 30.11.2023
Online: 24.04.2024
Keywords:
Sport psychology
Mental skills
Elite sport
High performance
Athlete
ISSN: 2703-240X
The purpose of this study was to examine New Zealand athletes’ attitudes towards sport psychologyconsultationasnosimilarstudyhasbeenconductedinNewZealandsince2004. Sixty-two New Zealand athletes (ranging from age-group to international-level), were administered the Sport Psychology Attitudes – Revised questionnaire (SPA-R). New Zealand athletes’ attitudes towards sport psychology have become more positive since 2004.Independentgroup t-testsdemonstrated thatathletes in2020 had significantly higher confidence in sport psychology, and significantly lower levels of personal openness and cultural preference comparedto athletes in2004 There wasno significant difference found in stigma tolerance scores. Chi-squared tests were carried out on nine different categories: gender, sport type (contact/non-contact and team/individual), previous exposure, age, current competitive level, top competitive level, number of individual and group sport psychology sessions, and employment level. Non-contact sport athletes were found to have significantly higher confidence than contact sport athletes and individual sport athletes were found to have significantly higher confidence and cultural preference than team sport athletes. A trend was found with higher level athletes having greater confidence in sport psychology and national level athletes were found to have higher cultural preference than junior athletes. These findings are useful for organisations and practitioners as they provide an understanding of athletes’ current perceptions and attitudes towards the field.
1. Introduction
For many, a relatively recent yet integral part of working towards optimal performance involves consulting with sport psychology consultants (SPC; Kornspan & Quartiroli, 2019). Sport psychology is defined as the psychological study of human behaviour in sport settings (Horn, 2008). In an applied sense, SPCs work with athletes and employ a variety of methods such as visualisation, self-talk, and mindfulness exercises to improve performance. Furthermore, SPCs look to engage with and address matters such as performance anxiety that can be detrimental to an athlete’s mental health and performance (Martin et al., 2004). The demand for and recognition of sport psychology has risen considerably, and many organisations are now realising the
benefits of deliberately working on the psychological aspects of performance (Rooney et al., 2021).
As of 2021, there are 26 accredited SPCs through High Performance Sport New Zealand (HPSNZ) and Sport and Exercise Science New Zealand (SESNZ). This number includes registered psychologists and accredited mental skills trainers. Because sport psychology is relatively new as a formal discipline when compared to other aspects of training, there is a lack of general awareness concerning its purpose and function from athletes, coaches, and the general public (Green et al., 2012). Furthermore, although consultation with practitioners has been shown to be beneficial for athletes’ performances and well-being (Kellmann et al., 2002), there can still be somewhat of a stigma and negative attitudes held towards sport psychology itself, as
*Corresponding Author: Sam Richardson, School of Sport Exercise and Nutrition, Massey University, New Zealand, Richardson.samuel.1146@gmail.com
well as towards those receiving sport psychology support (Green et al., 2012).
The aim of this study was to capture and understand New Zealand athletes’ current attitudes towards sport psychology. This is important as athletes’ attitudes towards the field will likely determine their intention, adherence, and future use of sport psychology (Martin, 2005). Moreover, understanding attitudes will also allow programmes and services to be refined to better support the well-being and performances of athletes and could help establish a more positive and stigma free environment (Lavallee et al., 2006; Martin et al., 2004). For example, if it is known that a perception exists within a group that engagement with a SPC is inherently revealing of mental health challenges, a consultant can look to address this and ideally reduce such stigma before initiating consultation.
Zakrajsek et al. (2023) recently updated their Multidimensional Model of Sport Psychology Service Provision (M2SP2-R), which identifies various elements that influence athlete attitudes, intentions to access, and engagement with sport psychology services, and, as such, can improve practitioner awareness of such variables. Athlete attitudes has been consistently reported in the literature (e.g., Martin et al., 2004; Zakrajsek et al., 2023) as playing an important role in influencing willingness to engage with sport psychology services and demonstrates the importance of understanding current trends.
Within the literature, a range of variables have been found to influence athletes’ attitudes towards sport psychology; specifically, (i) gender, (ii) sport type, (iii) nationality, (iv) previous exposure, and (v) age. For instance, it has been reported that males, along with contact sport (e.g., rugby, boxing) athletes, have generally held more negative attitudes towards sport psychology compared to females and non-contact sport (e.g., tennis, golf) athletes (Anderson et al., 2004; Martin, 2005; Martin et al., 2004; Wrisberg et al., 2009). It would seem this is due to ideas around masculinity and ‘macho’ cultures associated with being male and the kinds of sports that involve physical contact/collisions between athletes and higher risk of potential injury. Moreover, it seems that these environments often discouragethe expressionofemotion andsharingof problems and can develop a resistance to seeking help (Anderson et al., 2005; Martin, 2005).
Itappearsthattherearesomeculturalelementsanddifferences regarding the shaping of attitudes towards sport psychology. As of 2005, it appeared that, overall, New Zealand athletes held more favourable attitudes than athletes from Germany, United States, and Ireland (Lavallee et al., 2006). Overall, athletes that have had positive previous experiences had more favourable attitudes toward sport psychology consultation compared to those without experience or negative previous experiences (Anderson et al., 2004; Ildefonso et al., 2020; Martin et al., 2004; Martin, 2005; Wrisberg et al., 2009). Finally, the research regarding the impact of age is varied, with Martin (2005) finding difference between highschoolandcollegeathletes,whilst Andersonetal.(2004)and Shaw (2018) did not find any such differences. Overall, the literature shows that there are many aspects that can affect the attitudes of athletes towards sport psychology. It is important for organisations, coaches, and support staff and SPC themselves, to be aware of these aspects andpotential tendencies so that they can implementstrategieswithintheirprovisionofservicestomitigate,
as much as possible, likely challenges and, ultimately, provide the best possible support for the individuals in their care.
2. Methods
This research was assessed by the Massey University Human Ethics Committee and deemed as low risk (notification number: 4000023030).
2.1. Participants
A convenience sample of 62 New Zealand athletes (female n = 35, male n = 27) from five different age groups, 18 – 20 years (n = 30), 21 – 23 years (n = 18), 24 – 26 years (n = 4), 27 – 29 years (n = 4), and 30+ years (n = 6) were involved in this study. This study was exploratory and convenient in nature, therefore although small this sample size was deemed appropriate. The participants competed in a range of sports, including, cricket (n = 23), rugby (n = 13), athletics (n = 10), netball (n = 3), cycling (n = 2), soccer and hockey (n = 1), andother(n =9). Participantswereamixoftopinternational(n =1), international(n = 14), national (n = 34), junior(n = 8), and none (n = 5). Thirty-one participants (50%) had previously had an individual session with a SPC, and 47 participants (76%) had attended at least one sport psychology/mental skills workshop.
2.2. Questionnaire
To measure athletes’ attitudes towards sport psychology, the Sport Psychology Attitudes – Revised (SPA-R) was used. The SPA-R is a Likert scale questionnaire that was developed by Martin et al. (2002) to improve the validity and reliability of the Attitudes Towards Seeking Sport Psychology Consultation Questionnaire (ATSSPCQ; Martin et al., 1997) that had been primarily used from 1997 to 2002. Their analysis revealed factorial validity for use with a range of athletes (male/female; adolescent/adult) and has been used in various studies since. The SPA-R includes a 10-item demographics section to capture age, gender, level of sport, and previous exposure to sport psychology and mental skills. The remainder of the SPA-R consists of a fourfactor model involving a 7-point Likert scale for each of the four factors to determine an individual’s overall attitude towards seeking sport psychology consultation. These four factors are (i) stigma tolerance, (ii) confidence in sport psychology consulting, (iii) (lack of) personal openness, and (iv) cultural preference. The mean for each factor is determined by summing the scores and dividing by the number of items (e.g., an average score of higher than 5 for stigma tolerance illustrates the individual has concerns with the stigma associated with seeing a SPC). A high score in confidence in sport psychology illustrates the individual has high confidence in the field and believes it is useful. A high score in (lack of) personal openness indicates a lack of personal openness and unwillingness to share personal information. A high score in cultural preference indicates an individual would prefer a consultant of their own culture, race, or ethnicity.
2.3. Procedure
Key gatekeepers (i.e., coachers/managers) of various sporting organisations/teams were approached regarding the study, four of
which (Auckland Cricket, Harbour Rugby, Massey University AcademyofSport,HPSNZ)agreedtodistributealinktoanonline survey (carried out through Qualtrics.com) and information sheet to athletes via email. This link invited athletes to anonymously take part in a sport psychology attitudes questionnaire that would help deepen understanding of current attitudes towards sport psychology with the aim of utilising such insights to improve services in the future.
2.4. Statistical approach
All statistical analyses were performed with SPSS software for windows (IBM SPSS 311 Statistics 20, NY, USA). Descriptive values were obtained and reported as means and standard deviation (SD). Given that much of the data was categorical, and that Levene’s test and the Shapiro-Wilk Test provided > 50% significant data, data was analysed with a Chi-Squared Test and independent group t-tests. Validity for sample size was determined by checking against the result of Fisher’s Exact Test. Cohen’s d was calculated as a measure of effect size. Significance was accepted as p < 0.050.
3. Results
Table 1 compares the results of the current study from 2020 with those of Anderson et al. (2004) who conducted the last study of similar nature in New Zealand. These results show that in 2020, New Zealand athletes overall still hold positive attitudes towards sport psychology and still somewhat prefer working with SPCs of the same cultural background as themselves. Independent group t-tests demonstrate that athletes in 2020 have significantly higher confidence in sport psychology (t(141.74) = 2.911, p = 0.004) and lack of personal openness (t(137.56) = 4.855, p = 0.001); and significantly lower in cultural preference (t(118.03) = -2.942, p = 0.004). There was no significant difference found in stigma tolerance scores (t(128.74) = 1.043, p = 0.299).
As part of our analysis, gender, previous exposure, age, and highest competition level were also examined; however, no significant differences were found and therefore respective tables have not been included here.
Table 1: Stigma Tolerance (ST), Confidence in Sport Psychology Consulting (C), (Lack of) Personal Openness (PO), and Cultural Preference (CP) amongst cohorts of New Zealand athletes in the current study and Anderson et al. (2004).
significantly higher confidence in sport psychology and cultural preference compared to team sport athletes. Table 4 shows that there is a significant difference between athletes’ confidence in sport psychology at different competitive levels; national level athletes also had significantly higher cultural preference compared to junior athletes.
Table 2: Contact vs non-contact.
Scale Contact Non-contact
2.16 (1.10) 2.09 (1.33)
Notes: *p < 0.050.
Table 3: Team vs individual.
Notes: *p < 0.050.
Table 4: Current level.
(0.78)
(0.90)* 0.45
4.42 (0.90) 3.70 (1.00)* 0.76
3.43 (1.08) 3.92 (1.00)* 0.47
Notes: Values are mean (SD). *p < 0.010
As seen in Table 2, the findings suggest that non-contact sport athletes held significantly higher confidence levels than contact sport athletes. Looking at Table 3, individual sport athletes have
Notes: Top intl, Top international; Intl, Internationl; Nat’l, National; Jr, Junior *p < 0.050.
4. Discussion
The results from this study show that New Zealand athletes’ attitudes towards sport psychology consultation have become more favourable overall since 2004 (Table 1). Athletes have higherconfidenceinsportpsychologyandlessculturalpreference when working with a SPC. However, somewhat surprisingly, it is also important to note that over the last 16 years, athletes’ level of openness to sport psychology consultation has declined, and there was no significant change in perceived stigma from working with a sport psychologist.
Athlete confidence in sport psychology has likely improved since 2004 due to the increasing knowledge and research that has
been conducted regarding sport psychology in recent years (Kornspan & Quartiroli, 2019). Such research has provided new information, techniques, skills, and an increased awareness of the benefits of integrating sport psychology principles, including potential contributions towards athletes, coaches, and teams. It appears that this knowledge has filtered down from researchers and academics to SPCs and, finally, to coaches and athletes, which has improved confidence and overall integration. As a case in point, in recent years, HPSNZ has integrated education modules on sport psychology into various coach education programmes. Itis highlylikely that such workhas been improving awareness and attitudes and, as such, may be having a positive impact on normative and control beliefs which have been highlighted (e.g., Zakrajsek et al., 2023) as being important with regards to shaping overall attitudes and behaviours. Confidence in the field is considered as a key predictor of intention to utilise sport psychology services (Zakrajsek & Zizzi, 2007). Moreover, Anderson et al. (2004) found confidence to be the only of the four factors to significantly predict intention to engage with a SPC. Therefore, this improved confidence in sport psychology is an important development as it may lead to more athletes seeking proactive and/or remedial psychological support.
It would be expected this improved emphasis and integration would also improve athletes’ openness to sport psychology consultation. Furthermore, we have seen a major shift recently in the nature of the discourse around psychology and mental health, both in general populations as well as sport settings (Souter et al., 2018). Additionally, many sport organisations have improved accesstobothsupportandeducation.However,thisdoesnotseem to have yielded a significant shift in overall attitudes in New Zealand. In fact, athlete openness towards sport psychology consultation has decreased since 2004. It is important to acknowledge however that this finding could potentially be due to the high percentage (21%) of world class athletes in the study conducted by Anderson et al. compared to 1.6% in this current study. World class athletes typically have greater access and exposure to sport psychology support, often resulting to greater openness to consultation (Marin & Boone 1996). This higher proportion of world class athletes in the study conducted by Anderson et al. (2004) may have led to higher levels of personal openness due to greater exposure to sport psychology, compared to the current study where no significant correlation was found between previous exposure and attitudes.
The competitive level of athletes significantly impacted both their confidence and culturalpreference. No significant difference in confidence was found between specific competitive levels; however, an overall significant effect was found. Again, this is likely due to the small sample size for top international-level athletes in this study. Althoughthere was nosignificantdifference inconfidencebetweencompetitivelevels,atrendcanbeseenwith greater competitive levels having higher levels of confidence. MartinandBoone (1996)attributedasimilarfindingintheirstudy tohigherlevelathleteshavingmoreexposuretosportpsychology, and therefore greater appreciation and understanding of the importance of psychology. However, Anderson et al. (2004) examined competition level and attitudes towards sport psychology and found no significant differences.
The recorded decrease in cultural preference is likely found due to the increase in multiculturalism within New Zealand. As a country, New Zealand has a reputation as a modern and culturally
diverse country (Smits, 2011). Over the last two decades, multiculturalism and ethnic diversity within New Zealand has been increasing and is seen as one of the nation’s strengths (Simon-Kumar, 2020). Such growth may encourage New Zealanders to have more interactions and contact with individuals from other cultures, ethnicities, and backgrounds. Intercultural contact has been found to lead to higher levels of intercultural competence, which means individuals will have an improved ability and openness to communicate, function, and work effectively with people from other cultures (Schwarzenthal et al., 2020). Moreover, the importance of practitioners considering cultural elements and tailoring their delivery and interventions accordingly has been highlighted (e.g., Hodge et al., 2011) and may be having a positive impact on how the field is perceived. Therefore, athletes in 2020 would likely have lower levels of cultural preference compared to athletes of 2004.
Furthermore, this study also found non-contact sport athletes to have significantly higher confidence in sport psychology consultation compared to contact sport athletes. Similar results were found from Martin et al. (2004) and Martin (2005). It is believed that such attitudes are due to many contact sports (e.g., rugby, boxing) involving, and encouraging, aspects such as intimidation, toughness, and power, all of which are values commonly associated with masculinity and still prevalent inmany communities (Martin, 2005), including sport. Ultimately, it has been shown that environments that nurture such ways of thinking can reduce the likelihood of athletes holding positive attitudes towards sport psychology consultation and also nurture a perceived stigma surrounding help-seeking in general (Steinfeldt et al., 2009).
Differences were also found between athletes that compete in team sports and individual sports. The results indicated that individual sport athletes had higher confidence in sport psychology, whichhas been found inprevious work (e.g., Rooney et al., 2021). In their study, Rooney et al. attributed this difference to individual sport athletes having to rely exclusively on themselves and, therefore, engaging in greater psychological development in order to optimise performances and, as a result of such work, perceiving mental training as being beneficial (Rooney et al., 2021).
Interestingly, although team sport athletes reported lower levels of confidence in the field, findings suggested that these participants have lower levels of cultural preference (Table 3). Team sports are social practises where athletes are required to develop relationships and work with individuals of different ethnicities and cultures for the success of the overall team (Elling & Knoppers, 2005). As alluded to earlier, this would likely cause team sport athletes to develop higher levels of intercultural competence which may influence cultural preference (Schwarzenthal et al., 2020). National level athletes were found to have significantly lower cultural preference compared to junior athletes. According to Martin and Boone (1996), a lower cultural preference for national athletes would be expected because they found that attitudes towards sport psychology improved as competitive level gets higher. However, this also means that cultural preference should continue decreasing to international and top international level athletes, which was not found. This may have been due to this study having too few international and top international athletes to find an effect. It is important for
practitioners of sport psychology and mental skills to take these variables into account when working with athletes.
Although this study did have some significant and worthwhile findings, it is important to acknowledge some limitations. The most significant being the small sample size which can lead to false positives, as well as findings that are not representative of thetargetedpopulation.Thecurrentstudyalsoused aconvenience sample and, therefore, may not accurately reflect these sports and athletes. It is also important to note that this study also had a much higher percentage of younger athletes taking part with 77% of participants being aged 23 or under. Because age has the potential to affect athletes’ attitudes towards sport psychology (Martin, 2005), the findings from this study may not be generalisable to athletes over this age-group. Furthermore, the majority (81%) of athletes in this study were involved in team sports. Similarly to age, previous research (e.g., Rooney et al., 2021; Wrisberg et al., 2009) has shown that team and individual sport athletes overall have different attitudes surrounding sport psychology. Therefore, results from this study may be most appropriate to team sport environments.
In summary, this study re-examined and extended research on New Zealand athletes’ attitudes towards sport psychology from 16 years earlier by Anderson et al. (2004). It was found that New Zealand athletes still hold positive attitudes towards sport psychology overall and these attitudes have improved since 2004. New Zealand athletes were found to have low levels of stigma associated with seeking sport psychology consultation, high levels of confidence in the efficacy of sport psychology, moderate levels of personal openness and low to moderate levels of cultural preference. Both confidence in sport psychology and cultural preference have improved overall since 2004. However, although we have seen improvements, it seems as though personal openness has declined, and, as such, did not fit with the overall trend. This decrease in athletes’ personal openness is potentially due to some of this study’s limitations, such as the lack of international-level athletes and the relatively small sample size.
There is a limited amount of research with inconsistent findings regarding aspects that could affect an athlete’s attitudes towards sport psychology. Further research is needed to establish a more complete picture of such elements. For instance, this could include utilising a truly elite sample, and considering all variables within the M2SP2-R (Zakrajsek et al., 2023). This would provide a more complete understanding of current trends in attitudes, to inform programme development and ensure that our athletes are being provided with sport psychology services that they trust, as well as operating within environments where they feel safe to engage with such support.
Conflict of Interest
The authors declare no conflict of interests
References
Anderson, A. G., Hodge, K. P., Lavallee, D., & Martin, S. B. (2004). New Zealand athletes' attitudes towards seeking sport psychology consultation. The New Zealand Journal of Psychology, 33(3), 129–136.
Elling, A., & Knoppers, A. (2005). Sport, gender and ethnicity: Practises of symbolic inclusion/exclusion. Journal of Youth and Adolescence, 34(3), 257–268.
Green, M., Morgan, G., & Manley, A. (2012). Elite rugby league players' attitudes towards sport psychology consulting. Sport and Exercise Psychology Review, 8(1), 32–44.
Hodge, K., Sharp, L., & Heke, J. I. C. (2011). Sport psychology consulting with indigenous athletes: The case of New Zealand Māori. Journal of Clinical Sport Psychology, 5, 350–360.
Ildefonso, K., Blanton, J., Durwin, C., Arvinen-Barrow, M., & Kamphoff, C. (2020). A preliminary investigation into collegiate student-athletes’ attitudes towards athletic trainers and sport psychology consultants. Journal of Sport 8(1), 56–75.
Kornspan, A. S., & Quartiroli, A. (2019). A brief global history of sport psychology. American Psychological Association. Lavallee, D., Jennings, D., Anderson, A. G., & Martin, S. B. (2006). Irish athletes’ attitudes toward seeking sport psychology consultation. The Irish Journal of Psychology, 26(3–4), 115–121.
Martin, S. (2005). High school and college athletes' attitudes toward sport psychology consulting. Journal of Applied Sport Psychology, 17(2), 127–139.
Martin, S. B., & Boone, K. (1996). NCAA Division 1 football players' attitudes toward seeking sport psychology consultation. Journal of Applied Sport Psychology, 8, S128. Martin, S. B., Kellmann, M., Lavallee, D., & Page, S. J. (2002). Development and psychometric evaluation of the sport psychology attitudes revised form: A multiple group investigation. The Sport Psychologist, 16(3), 272–290.
Martin, S. B., Lavallee, D., Kellmann, M., & Page, S. J. (2004). Attitudes toward sport psychology consulting of adult athletes from the United States, United Kingdom, and Germany. International Journal of Sport and Exercise Psychology, 2(2), 146–160.
Martin, S. B., Wrisberg, C. A., Beitel, P. A., & Lounsbury, J. (1997). NCAA Division I athletes’ attitudes toward seeking sport psychology consultation: The development of an objective instrument. TheSport Psychologist, 11(2), 201–218.
Rooney, D., Heron, N., & Jackson, R. (2021). Attitudes towards sport psychology consulting in athletes from individual and team sports. BMCSportScience,MedicineandRehabilitation, 13(1), 1–7.
Schwarzenthal, M., Schachner, M. K., Juang, L. P., & van de Vijver, F. J. R. (2020). Reaping the benefits of cultural diversity: Classroom cultural diversity climate and students' intercultural competence. European Journal of Social Psychology, 50(2), 323–346.
Shaw, A. R. (2018). What's the mindset? An investigation of junior hockey players' attitudes toward sport psychology [Unpublished master’s thesis]. University of Lethbridge. Simon-Kumar, R. (2020). Intersections of inequality, migration and diversification. Springer.
Smits, K. (2011). Justifying multiculturalism: Social justice, diversity and national identity in Australia and New Zealand. Australian Journal of Political Science, 46(1), 87–103.
Souter, G., Lewis, R., & Serrant, L. (2018). Men, mental health and elite sport: A narrative review. Sports Medicine-Open, 4(1), 1–8.
Richardson et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 1-6 (2024)
Steinfeldt, J. A., Steinfeldt, M. C., England, B. & Speight, Q. (2009). Gender role conflict and stigma toward help-seeking among college football players. Psychology of Men & Masculinity, 10(4), 260–270. Wrisberg, C. A., Simpson, D., Loberg, L. A., Withycombe, J. L., & Reed, A. (2009). NCAA division-I student-athletes’ receptivity to mental skills training by sport psychology consultants. The Sport Psychologist, 23(4), 470–486.
Zakrajsek, R. A., Martin, S. B., Thomson, S. R., & Gulliver, A. (2023). Stigmatisation of Mental Illness and Seeking Sport
Psychology Services. In I. Nixdorf, R. Nixdorf, J. Beckmann, S. Martin, & T. Macintyre (Eds.), Routledge Handbook of Mental Health in Elite Sport (pp. 267–286). Routledge. Zakrajsek, R. A., & Zizzi, S. J. (2007). Factors influencing track and swimming coaches’ intentions to use sport psychology services. Athlete Insight: Online Journal of Sport Psychology, 9(2),1–21. Retrieved from https://researchrepository.wvu.edu/faculty_publications/2922 /
The Journal of Sport and Exercise Science, Vol. 8, Issue 1, 7-13 (2024)
www.jses.net
Perceptionof body image,useofdietarysupplements, and doping among malegym trainers in Sri Lanka
Maramba Vidanage Chandima Madushani 1 , Thalagalage Shalika Harshani Perera 1,2 , Ajmol Ali 2 , David Stephen Rowlands2* , Dingirige Sameera Lakruwan Perera3
1Department of Sport Sciences and Physical Education, Faculty of Applied Sciences, Sabaragamuwa University of Sri Lanka, Sri Lanka
2School of Sport, Exercise and Nutrition, Collage of Health, Massey University, New Zealand
3Department of Sports Science, Faculty of Applied Sciences, University of Sri Jayewardenepura, Sri Lanka
A R T I C L E I N F O
Received: 10.03.2022
Accepted: 05.11.2023
Online: 25.04.2023
Keywords:
Supplements
MaleBodyAttitudeScale
PerformanceEnhancingAttitudeScale
Lenientattitude
Muscularity
Bodyfat
A B S T R A C T
People who train in public gymnasiums are motivated to improve aesthetic appearance, performance, and health, which may sometimes include the use of dietary supplements or banned substances. Accordingly, gym trainers are ideally placed to provide information, but the relationship between dietary supplementation and doping use and attitudes towards body image in Sri Lankan gym trainers is still being determined. 150 National Vocational Qualification certified male gym trainers across all of Sri Lanka were recruited into the study Data were gathered to analyse supplement use, and the Male Body Attitude Scale (MBAS) and Performance Enhancing Attitude Scale (PEAS) tools were used to analyse body image and attitudes to doping, respectively. Gym trainers had positive attitudes toward body image and were satisfied with their muscularity (mean = 2.6, SD = 0.1) and body fat levels (mean = 2.5, SD = 0.1). All participants reported using at least one dietary supplement, caffeine-containing beverages (relative frequency 90%), protein powders (49%), herbal supplements (41%), individual vitamins (35%), individual minerals (34%) andenergydrinks(25%).Half(54%)ofgymtrainershadalenientpositive attitudetowards dopingthat wasassociatedwith bodyimage.Supplementuseandperceptionofmuscularity (r = 0.55, p < 0.001) and body fat (r = 0.23, p = 0.011) were positively correlated. SignificantregressionassociationsexistedbetweenPEASand muscularity,bodyfat,height, and supplement use. Gym trainers had a high frequency of supplement use, and a lenient attitude towards doping, which is associated with a positive perception of body composition. Confirmation of attitudinal transference to clients requires further research.
1. Introduction
Sufficient evidence supports the idea that people are motivated to exercise in gymnasiums for reasons including maintaining body image and composition, physical fitness, sports performance, metabolic and mental health (Lamarche et al., 2018). However, coaching is often sought, and gym trainers play a crucial role in assisting clients with theirgoals byprescribingschedules, training
programs, and nutritional guidance. Additionally, trainers provide support and training structures and recommendations for athletes. With respect to health, fitness, and as a determinant of performance in some sports, body composition is used to describe the percentages of fat and fat-free mass, including bone and muscle in the body (Campa et al., 2021). This perception can be positive or negative and is influenced by personal and environmental factors, such as individual perception, feelings,
*Corresponding Author: David S. Rowlands, School of Sport, Exercise, and Nutrition, Massey University, New Zealand, d.s.rowlands@massey.ac.nz
and associated behaviours (Campa et al., 2021). The perception of body image has been found to significantly impact behavioral, cognitive, and affective outcomes, potentially influencing a person's quality of life (Pruzinsky et al., 2002). Body image concerns often fall between satisfaction and dissatisfaction with physical appearance(Thompson, 2004). Body dissatisfaction may lead to the implementation of harmful weight change plans, including unnecessary exercise, weight loss, and the use of supplements(McCabe & Ricciardelli, 2003).
Dietary supplements are defined as products (other than tobacco) intended to supplement the diet and contain one or more dietary ingredients. Examples of dietary supplements include vitamins, minerals, herbs, and amino acids, which can be consumed in various forms, such as tablets, capsules, powders, and drinks(Jawadi et al., 2017). The use of dietary supplements is prevalent among individuals who attend both commercial and non-commercial gyms(Morrison et al., 2004). People involved in physicalorathleticactivitiestendtousedietarysupplementsmore frequently than others to increase or maintain muscle mass, strength, power, exercise recovery, performance, and weight control (Attlee et al., 2017; Caudwell & Keatley, 2016; Khoury & Antoine‐Jonville, 2012). However, some supplements may have side effects and can be harmful (Jenkinson & Harbert, 2008), while others may contain banned substances, leading to positive doping outcomes (Maughan et al., 2004, Morente-Sánchez & Zabala, 2013). A doping attitude means an individual’s predisposition toward using performance-enhancing substances and methods (Baron et al., 2007), and if athletes use illegal substances to enhance performance, it is known as doping (Brand et al., 2014). In a study by Ruano and Teixeira (2020), the most preferred sources of information regarding supplements were registered dietitians, the internet, fitness coaches, friends, and pharmacists.
Previous research has mainly focused on investigating the prevalence of supplement use and the relationship between doping and body dissatisfaction among elite athletes, university students, and adolescents (Backhouse et al., 2013, Bloodworth et al., 2012, Hildebrandt et al., 2012) Therefore, this study aims to examine the perception of body image, dietary supplement use, and attitude towards doping among Nationally Qualified male gym trainers in Sri Lanka. We hypothesise that there will be a positive relationship between the negative perception of body image among gym trainers in Sri Lanka and supplement use, and we also aim to explore potential differences based on age.
2. Methods
During the sampling period (20/06/2020 to 10/07/2020), 300 registered National Vocational Qualification certified male gym trainers were at the National Apprentice and Industrial Training Authority, Sri Lanka. A systematic sampling technique was used to select and invite 150 gym trainers from this population. The protocol was approved by the Sabaragamuwa University of Sri Lanka ethics committee, and written informed consent was given by the participants.
A single-sample age-stratified study design was used to examine theperception of body image,use ofdietarysupplements, and attitude towards banned substances. Invitations to participate were sent via email, and a questionnaire and consent form were
also sent and collected via email. The survey questionnaires obtained personal information (age, height, weight), supplement usage, and questions related to the Male Body Attitude Scale (MBAS) (Tylka et al., 2005) and the Performance Enhancing AttitudeScale(PEAS)(Moran,etal.,2008). HigherMBASscores reflect a more negative body attitude, and lower scores indicate a more positive body attitude. The MBAS provides measures of attitude toward muscularity, low body fat, and height and comprises of 29 questions, with the MBAS total score ranging from 29 to 174. Body dissatisfaction (MBAS) was measured using a questionnaire via a 6-point Likert scale: 1-never, 2-rarely, 3-sometimes, 4-often, 5-usually, and 6-always.
Attitudetowardsdopingwasmeasured usingthePEAS,which is an instrument of unidimensional self-reports that measures general attitude towards doping. It is an extensively used questionnaire to assess doping attitudes among adult and adolescent athletes. It is widely used in doping literature(Nicholls et al., 2017) to explore the relationship between attitudes to doping and supplement use perfectionism, achievement goals and the motivational climate, willingness to dope, and social desirability. Some of these samples have included adults, teenagers, or a mix of older and younger adults (Nicholls et al., 2017). PEAS comprises 17 items measured on a 6-point Likert scale: 1-strongly agree, 2-agree, 3-slightly agree, 4-slightly disagree, 5-disagree, and 6-strongly disagree. The PEAS total score ranges from 17 to 102, and the PEAS theoretical middle and the neutral point is 59.5; a higher than neutral suggests a positive attitude towards doping(Hildebrandt et al., 2012) indicating more support for doping in s sports.
All data gathered from the survey was entered into a spreadsheet prior to analysis (SPSS version 21.0, Chicago, IL, USA). One-way ANOVA tests were performed to test for significant differences in body image regarding age, with p < 0.050 taken as significant. A linear regression analysis (see Supplemental materials) was performed to estimate the impact of variables on doping attitude. Pearson correlations were performed on the relationship between dietary supplement use and body image parameters. The strength of the relationship of the correlation was interpreted using the following threshold: 0 to 0.1 as trivial; 0.1 to 0.3 as small; 0.3 to 0.5 as moderate; 0.5 to 0.7 as large; 0.7 to 0.9 very large; and greater than 0.9 as nearly perfect (Hopkins et al., 2009)
3. Results
All 150 invitees agreed to participate in the study. Participants' mean weight, height, and BMI were 73.6 kg (SD = 9.6), 171.7 cm (SD = 5.6), and 24.9 kg/m² (SD = 2.5), respectively. The number of participants (n) and frequency (%) within each age categorisation in years was: < 20 years (n = 1, 0.7%), 20 – 29 years (n = 105, 70%), 30 – 39 years (n = 39, 26%), and ≥ 40 years (n = 5, 3.3%).
3.1. Sources of information
Most of the gym trainers gained information from the Internet (75%; all significantly different from other sources, p < 0.050), followed by information from friends (22%), newspapers (2%), and doctors (1%).
3.2. Perception of body image
Figure 1 illustrates the mean sample response in the MBAS subscales (muscularity, body fat, height, overall body). The highest scores were recorded with height, reflecting more negative body-composition attitudes, and the lower scores in muscularity and body fat reflected more positive attitudes towards the body composition parameters. The overall mean score for MBAS across the group was 2.5.
Figure 1: Male Body Attitude Scale (MBAS) scores. Data shows means and error bars represent standard deviations.
3.3. Supplement use
All participants reported using at least one dietary supplement, withthetypeandfrequency ofuseshowninTable1. Mosttrainers used caffeinated beverages (tea, coffee, herbal tea), at a ratio of use nearly two-foldgreaterthan othersupplements. The leastused supplement was caffeinated candy/gum/medications.
Table 1: Type and frequency of supplement use in Sri Lankan male gym trainers. Supplement
Caffeinated-containing beverages (e.g., instant coffee, hot brewed tea)
Protein powder (e.g., whey, soy)
Plant extracts/herbal supplements (e.g., echinacea, ginseng)
Individual vitamins (e.g., vitamins A, C, D, E)
Individual minerals (e.g., calcium, iron)
3.4. Attitude toward doping
The mean sample total PEAS score was 64.4 (SD = 28.1), suggesting that gym trainers have a slight, somewhat positive attitude towards doping (neutral point 59.5; standardised difference relative to the neutral point 0.17). The percentage of gym trainers scoring higher than mid-value was 54%.
3.5. Relationshipbetweenbodyimageanddietarysupplementuse
The correlation relationships between supplement use and MBAS categories and PEAS are shown in Table 2. Small correlations were observed between individual vitamins (correlation: negative), vitamins and minerals (negative), caffeine-containing beverages (positive), and branched-chain amino acids (negative) and PEAS. No significant correlations were observed with muscularity. Small correlations were observed between plant extracts/herbal supplements (negative) and muscularity; individual vitamins, plant extracts/herbal supplements and height (negative); and caffeinated beverages and body fat (positive). The correlations between supplement use and perception of muscularity and body fat were 0.55 and 0.23, respectively.
Note: *p < 0.050, **p < 0.001.
Table 2: Correlations between supplement use and MBAS and PEAS
3.6. Age and the relationship between body image and attitudes to doping
Figure 2 displays the mean response in overall MBAS and PEAS scores relative to age, and Table 3 the age-contrast statistics. For MBAS, 20 – 29 years was significantly lower than 30 – 40 years, but no other significant contrasts existed. A regression analysis was performed to predict PEAS from age, overall body image, and supplement use. The outcome was: PEAS = -0.05644 + 0.0236 × (Age) + 0.492 × (Overall Body image) – 0.270 × (Supplement use). The R-squared value for the fitted model indicates that 63.2% ofthevariation in the PEAS canbe explained by age, overall body, and supplement use, but overall body image was the greatest predictor
Figure 2: Group response to (A) Male Body Attitude Scale (MBAS) and (B) Performance Enhancing Attitude Scale (PEAS)
Sample size in each age band:20 – 29 years, n =87; 30 – 40years, n = 33; ≥ 40 years, n = 5. Age band > 20 years was excluded due to low sample bias (n = 1). Data shows means and error bars represent standard deviations
Table 3: The effect of age on the PEAS and the MBAS
Age band (years)
PEAS
Note: PEAS, Performance Enhancing Attitude Scale; MBAS, Male Body Activity Scale; LL, lower limit; UL, upper limit. Age band > 20 years was excluded due to low sample bias (n = 1).
4. Discussion
The current study examined the perception of body image, use of dietary supplements, and attitudes toward doping among nationally qualified male gym trainers in Sri Lanka. Our findings indicate that all gym trainers used supplements; however, this included tea, coffee, and herbal tea, which some may argue are not traditionally considered supplements. Additionally, the available evidence suggests that most gym trainers had abovenormal BMI values.
Based on evidence regarding the average MBAS score, it seems fair to suggest that the gym trainers were overall satisfied with their bodies. However, gym trainers were dissatisfied with their height (most think they need to be taller) and were satisfied with muscularity and body fat. On logical grounds, there is no compelling reason to argue that these findings suggest the fact it may be that gym trainers are satisfied with the subscales which they have control via training (muscularity, body fat) and dissatisfied with the subscale with they cannot change (height). Similarly, in adolescent boys Yager and O’Dea(Yager & O’Dea, 2014) found in response to direct questions about body image and body dissatisfaction, 78% indicated that they were satisfied with their body muscularity, fat, and height, while 10% thought themselves too thin and 13% too fat. In contrast, for the question about body dissatisfaction, 35%, 30%, and 28% of the boys indicated that they would like to be their present weight, a little lighter, or a little heavier, respectively.
In the present study, all participants used at least one dietary supplement, with caffeine-containing drinks, then protein powders, the top-ranked items. Ruano and Teixeira (Ruano & Teixeira, 2020) reported that 44% of gym trainers used dietary supplements (Morente-Sánchez & Zabala, 2013). According to that study, the most consumed supplements were protein powders (80%), followed by multivitamins and/or minerals (38%), sports bars (37%), branched-chain amino acids (37%), and n-3 fatty acids (36%)(Ruano & Teixeira, 2020). Gym trainers in our study were found to be frequent consumers of supplements that may be associated with performance-enhancing motives, such asvitamins, minerals, branched-chain amino acids, and caffeine-containing drinks (Table 2).
Our study revealed that over half (54%) of Sri Lankan gym trainers had PEAS scores higher than the theoretical value, indicating a lenient attitude toward doping. This finding is consistent with the study by Yager and O’Dea (2014), which found a positive attitude toward doping among Australian adolescent boys. However, Sas-Nowosielski and Budzisz (2018) observed anegativeattitude toward dopingamongPolish athletes. These contrasting results suggest that attitudes toward doping may vary across different populations.
Yager and O’Dea (2014) found a relationship between supplement use and subscales of MBAS, which implies a positive relationship between supplement use and muscularity, body fat, and height. In this study, only plant and herbal supplement use in gym trainers had a small association with the perception of muscularity, with only caffeine associated with body fat. Height is largely determined during childhood and adolescent growth phases, and while plant and herbal supplement and vitamin use had a small negative correlation with height in adult gym trainers, there is unlikely to be any biological relationship.
Yager and O’Dea (2014) also stated that body image and dissatisfaction levels vary according to age. They stated that total MBAS scores and dissatisfaction with muscularity increased with age. Males older than 16 years of age were significantly more likely to have higher scores on the muscularity subscale and the total MBAS indicating greater levels of body dissatisfaction. The current study shows that body image varied according to age only between 20 – 29 years and 30 – 40 years (Table 3). Similarly, we found no significant difference in mean PEAS scores by age, suggesting performance-enhancing attitudes to doping attitudes were generalisable.
In the current study, most gym trainers received information from the Internet (75%); only 1% received advice from doctors. It is well known that information from the internet is of variable quality. The study gathered no data on the quality of information, but future studies could explore the effect of information quality on attitudes. In contrast, Waddington et al. (2005) stated that 28% of English professional footballers took advice from the club’s physiotherapist, 21% from a fitness trainer, 21% from another sports scientist (e.g., nutritionist), and the club’s doctor was their last option (15%). The differences may reflect variations in participant cohorts or the availability of sports and sports medical professionals in Sri Lanka to provide advice.
This study shed light on the relationship between body image, use of dietary supplements, and attitudes toward doping in sports. Strengths of the study include a sizable sample and the use of standardised measures to assess body image and attitudes toward doping in sports. However, a limitation is that some of the words and concepts in the questionnaire may have needed to have been correctly understood by gym trainers, which could have led to erroneous reporting.
Conclusion
The findings suggest that gym trainers had a positive attitude toward body image and were satisfied with their muscularity and body fat levels but not with their height. The study revealed that all participants reported using at least one dietary supplement, with the most used supplements being caffeine-containing beverages, energy drinks, protein powders, herbal supplements,
individual vitamins, protein bars and individual minerals. A slight majority of gym trainers exhibited a lenient attitude toward doping, which was significantly associated with body image. Most gym trainers obtain information from the Internet.
Based on these findings, it is important to emphasize the need for reliable information for self-learning, considering the risks of side effects from various supplements and the potential influence of trainer attitudes on clients, leading to doping abuse in sports. Further research in this area may involve conducting similar studies with gymusers toexamine theirperceptionofbodyimage, supplement use, and attitude toward doping, as well as the sources and influences of information, including coaches and gym staff.
Conflict of Interest
The authors declare no conflict of interest
Acknowledgment
We thank the participants for their contributions. The authors declare no conflicts of interest. Author contributions: MVCM, TSHP, and DSLP conceived and conducted the study, analysed data, and drafted the manuscript. AA and DSR contributed to data analysis, interpretation, manuscript writing, and supervision. All authors read and approved the final version of the manuscript.
References
Attlee, A., Haider, A., Hassan, A., Alzamil, N., Hashim, M., & Obaid, R. S. (2018). Dietary supplement intake and associated factors among gym users in a university community. Journal of Dietary Supplements, 15(1), 88–97. https://doi.org/10.1080/19390211.2017.1326430
Axon, D. R., Vanova, J., Edel, C., & Slack, M. (2017). Dietary supplement use, knowledge, and perceptions among student pharmacists. American Journal of Pharmaceutical Education, 81(5),1–12. https://doi.org/10.5688/ajpe81592
Backhouse, S. H., Whitaker, L., & Petróczi, A. (2013). Gateway to doping? Supplement use in the context of preferred competitive situations, doping attitude, beliefs, and norms. Scandinavian Journal of Medicine & Science in Sports, 23(2), 244–252. https://doi.org/10.1111/j.1600-0838.2011.01374.x
Baron, D. A., Martin, D. M., & Abol Magd, S. (2007). Doping in sports and its spread to at-risk populations: An international review. World psychiatry: Official Journal of the World Psychiatric Association (WPA), 6(2), 118–123.
Bloodworth, A. J., Petróczi, A., Bailey, R., Pearce, G., & McNamee, M. J. (2012). Doping and supplementation: The attitudes of talented young athletes. Scandinavian Journal of Medicine & Science in Sports, 22(2), 293–301. https://doi.org/10.1111/j.1600-0838.2010.01239.x
Brand, R., Heck, P., & Ziegler, M. (2014). Illegal performance enhancing drugs and doping in sport: A picture-based brief implicit association test for measuring athletes’ attitudes. Substance Abuse Treatment, Prevention, and Policy, 9(1). 1–11. https://doi.org/10.1186/1747-597x-9-7
Campa, F., Toselli, S., Mazzilli, M., Gobbo, L. A., & Coratella, G. (2021). Assessment of body composition in athletes: A narrative review of available methods with special reference
Madushani et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 7-13 (2024)
to quantitative and qualitative bioimpedance analysis. Nutrients, 13(5), 1–25. https://doi.org/10.3390/nu13051620
Caudwell, K. M., & Keatley, D. A. (2016). The effect of men’s body attitudes and motivation for gym attendance. Journal of Strength and Conditioning Research, 30(9), 2550–2556. https://doi.org/10.1519/jsc.0000000000001344
ElKhoury,D.,&Antoine-Jonville,S.(2012).Intakeofnutritional supplements among people exercising in gyms in Beirut City. Journal of Nutrition and Metabolism, 2012, 1–12. https://doi.org/10.1155/2012/703490
Hildebrandt, T., Harty, S., & Langenbucher, J. W. (2012). Fitness supplements as a gateway substance for anabolic-androgenic steroid use. Psychology of Addictive Behaviors, 26(4), 955–962. https://doi.org/10.1037/a0027877
Hopkings, W. G., Marshall, S. W., Batterham, A. M., & HANIN, J. (2009). Progressive statistics for studies in Sports Medicine and Exercise Science. Medicine & Science in Sports & Exercise, 41(1), 3–12.
Jawadi, A. H., Addar, A. M., Alazzam, A. S., Alrabieah, F. O., Al Alsheikh, A. S., Amer, R. R., Aldrees, A. A., Al Turki, M. A., Osman, A. K., & Badri, M. (2017). Prevalence of dietary supplements use among gymnasium users. Journal of Nutrition and Metabolism, 2017, 1–8. https://doi.org/10.1155/2017/9219361
Jenkinson, D. M., & Harbert, A. J. (2008). Supplements and sports. American Family Physician, 78(9), 1039–1046. Lamarche, L., Gammage, K. L., & Ozimok, B. (2018). The gym as a culture of body achievement: Exploring negative and positive body image experiences in men attending university. Sage Open, 8(2), 1–12. 215824401877810. https://doi.org/10.1177/2158244018778103
Maughan, R. J., King, D. S., & Lea, T. (2004). Dietary supplements. Journal of Sports Sciences, 22(1), 95–113. https://doi.org/10.1080/0264041031000140581
McCabe, M. P., & Ricciardelli, L. A. (2003). Body image and strategies to lose weight and increase muscle among boys and girls. Health Psychology, 22(1), 39–46. https://doi.org/10.1037/0278-6133.22.1.39
Moran, A., Guerin, S., Kirby, K. (2023, October 3) The development and validation of a doping attitudes and behaviour scale (DABS). World Anti-Doping Agency. https://www.wada-ama.org/en/resources/social-scienceresearch/development-and-validation-doping-attitudes-andbehaviour-scale
Morente-Sánchez, J., & Zabala, M. (2013). Doping in sport: A review of elite athletes’ attitudes, beliefs, and knowledge. Sports Medicine, 43(6), 395–411. https://doi.org/10.1007/s40279-013-0037-x
Morrison, L. J., Gizis, F., & Shorter, B. (2004). Prevalent use of dietary supplements among people who exercise at a commercial gym. InternationalJournalofSportNutritionand Exercise Metabolism, 14(4), 481–492. https://doi.org/10.1123/ijsnem.14.4.481
Nicholls, A. R., Madigan, D. J., & Levy, A. R. (2017). A confirmatory factor analysis of the Performance Enhancement Attitude Scale for adult and adolescent athletes. Psychologyof Sport and Exercise, 28, 100–104. https://doi.org/10.1016/j.psychsport.2016.10.010
Pruzinsky, T., & Cash, T. F. (2002). Understanding body images: Historical and contemporary perspectives. In T. F. Cash, & T.
Pruzinsky(Eds.), Bodyimage:Ahandbookoftheory,research, and clinical practice (pp. 3–12). Guilford Press.
Ruano, J., & Teixeira, V. H. (2020). Prevalence of dietary supplement use by gym members in Portugal and associated factors. Journal of the International Society of Sports Nutrition, 17(1). 1–8. https://doi.org/10.1186/s12970-02000342-z
Sas-Nowosielski, K., & Budzisz, A. (2018). Attitudes toward doping among Polish athletes measured with the Polish version of Petroczi’s Performance Enhancement Attitude Scale. Polish Journal of Sport and Tourism, 25(2), 10–13. https://doi.org/10.2478/pjst-2018-0008
Scofield, D. E., & Unruh, S. (2006). Dietary supplement use among adolescent athletes in central Nebraska and their sources of information. The Journal of Strength and Conditioning Research, 20(2), 452–455. https://doi.org/10.1519/r-16984.1
Thompson, J. K. (2004). The (mis)measurement of body image: Ten strategies to improve assessment for applied and research purposes. Body Image, 1(1), 7–14. https://doi.org/10.1016/s1740-1445(03)00004-4
Tylka, T. L., Bergeron, D., & Schwartz, J. P. (2005). Development and psychometric evaluation of the Male Body Attitudes Scale (MBAS). Body Image, 2(2), 161–175. https://doi.org/10.1016/j.bodyim.2005.03.001
Waddington, I. (2005). Drug use in English professional football. British Journal of Sports Medicine, 39(4), 211 https://doi.org/10.1136/bjsm.2004.012468
Yager, Z., & O’Dea, J. A. (2014). Relationships between body image, nutritional supplement use, and attitudes towards doping in sport among adolescent boys: Implications for prevention programs. Journal of the International Society of Sports Nutrition, 11(1). 11–13. https://doi.org/10.1186/15502783-11-13
Supplemental materials
Supplementary Table 1: Regression statistics.
Supplementary Table 2: ANOVA statistics.
Madushani et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 7-13 (2024)
Supplementary Table 3: Regression coefficients and equation.
Note: PEAS = -0.0252 ‒ 0.0350 × (Muscularity) + 0.1762 × (Body Fat) + 0.0722 × (Height) + 0.7890 × (Supplement use)
The Journal of Sport and Exercise Science, Vol. 8, Issue 1, 14-19 (2024)
www.jses.net
Career entry and early experiences of sport scientists in Australia
Lyndell
Bruce1* , Brad Aisbett2 , Peter Kremer1
1CentreforSportResearchwithinInstituteofPhysicalActivityandNutrition,SchoolofExerciseandNutritionSciences,DeakinUniversity, Australia
2Institute of Physical Activity and Nutrition, School of Exercise and Nutrition Sciences, Deakin University, Australia A R T I C L E I N F O
Received: 09.06.2023
Accepted: 09.07.2024
Online: 16.09.2024
Keywords:
Volunteer roles
Internship
Sport science
Work integrated learning
Placement
1. Introduction
Little is understood about the early career experiences of sport scientists in Australia. The aim of this manuscript was to investigate the reasons why people pursue a career in sport science, alongside their early career experiences, including those experiences that early career sport scientists found to be valuable for their current role. 116 Australian Sport Scientists completed an online survey aimed at understanding why they entered the profession, and their early career experiences including the number of paid and unpaid intern (or similar) roles (excluding work integrated learning) they had before their current (paid) position. Descriptive statistics revealed that participants pursued a sport science careerto followtheirpassionforsport andbecause italignedwith theirperceived abilities. Sport scientists who were employed reported a median of 3 paid and unpaid roles before obtaining their first paid role, while those who were not currently employed reported a median of 5 paid and unpaid roles to this point of their career. Internship positions and on the job training were considered the most helpful activities that assisted early in their career. The requirement of internships beyond work integrated learning gained through formal university study suggests there is a gap between knowledge and /or skills and what is required for employment, or the time required to refine their practices.
The term ‘sport science’ is broad and can be used to describe a wide range of potential roles for people working in sport (e.g., sports physiologist, biomechanist, skill acquisition specialist, performance analyst, and strength andconditioning coach; French & Torres-Ronda, 2022). Some roles within the sport science industry are specialist, whereby the employee has a sole focus in one discipline area or generalist, whereby an employee may undertake roles across more than one discipline. Many roles within the industry are in high performance sport, and the opportunity to work in this field appeals to many young sports fans. At least 34 Australian universities offer studies within exercise and sport science, with many offering more than one course option (Exercise and Sports Science Australia, 2022a). Currently, little is known about why people decide to pursue a career in sport science.
Previous research has begun to explore the reasons why students enter an exercise and sport science degree and career. Spittle et al. (2021) investigated a sample of Australian undergraduate exercise and sport science students at one Australian university and asked them why they chose to study exercise and sport science. The strongest reason for pursuing a career in exercise and sport science was related to ‘sport association and continuation’, followed by ‘interpersonal reasons’, ‘means to an end’, and ‘subjective warrant and prestige’. York et al. (2014) interviewed six sport scientists and three indicated they initially had little understanding of the role. Given thebreadthofcareeroptionsavailablefromexerciseandsport science courses, these limited findings may not accurately reflect those of students and graduates who have specifically undertaken an exercise and sport science course to pursue a career in sport science (rather than using the course as a pathway to another vocational outcome). Understanding motivations for entering the sport science workforce could provide valuable information for universities and career educators to use in promotion and career guidance.
*Corresponding Author: Lyndell Bruce, Deakin University, Australia, lyndell.bruce@deakin.edu.au
Entryintoanyworkforcecanbechallenging,especiallywhenthe number of graduates is greater than the demand. Anecdotally, entry into the sport science workforce, especially in the high-performance setting, is very challenging. Stevens et al. (2018) found that 41% of exercise and sport science graduates had to volunteer, in addition to their formal work integrated learning (WIL) placements before becoming paid in the exercise and sport science workforce. This ranged from short term (7 to 31 days, 5% of the workforce) to long term (> 12 months, 18% of the workforce). Currently there is no evidence of how sport science graduates enter the workforce, and what volume of volunteer or paid internship roles they undertake beforepotentiallygainingfull-timeemployment(eitherinafull-time capacity or multiple part-time equivalents). Understanding how graduates enter the workforce and the experiences gained in the process will assist University sport science course designers. It will also assist in providing career advice to future sport scientists and policy developers around acceptable WIL and internship practices.
The number of accredited sport scientists in Australia has risen from 5 in 2008 (0.25% of the membership base of Exercise and Sports Science Australia [ESSA], the accrediting body within Australia)to398plus84highperformancemanagersin2021(5.8%; Exercise and Sports Science Australia, 2021). This is in part due to the increased professionalism in women’s sport (Bowes & Culvin, 2021; McLachlan, 2019). Two separate, survey-based, studies have reported on the sport science workforce in Australia–one was conducted in 2013, producing a report (Dawson et al., 2013) and a publication(Dwyeretal.,2019),andanotherwasconductedin2019, producing a publication (Bruce et al., 2022). These enable comparisons of the demographic profiles of participants across the two occasions. Notable profile differences include: a greater proportion of females working in the industry in 2019 compared to 2013(2013=24%;2019=33%);agreaterproportionofparticipants reportedworkingintwoormorepositionsin2019comparedto2013 (2013 = 23%; 2019 = 44%); and a greater proportion of participants had been in the sport science industry for more than five years in 2019 compared to 2013 (2013 = 37%; 2019 = 62%; Bruce et al., 2022; Dawson et al., 2013). Furthermore, a higher proportion of participantshadcompletedahigherlevelofeducation(Masters/PhD) in 2019 (68%) than in 2013 (57%). These figures help understand the evolving Australian sport science industry and how it may be changing.
As the sport science workforce continues to grow and evolve, it is important to understand what factors sport scientists believe assisted them in gaining employment. Stevens et al. (2021) asked sport science practitioners what their main tasks were, with participants required to select three tasks from a list of 14, with the option to nominate other specific tasks. Four tasks received greater thanathirdofpreferences:assessmentoffitness/performance(56%), training monitoring (53%), designing, implementing, and modifying training programs (42%), and research (42%). This provides some understanding of the tasks a sport scientist may be required to complete but does not provide any information about whether sport scientists feel well-prepared to perform this work.
The aim of the present study was to investigate why people enter the sport science workforce. In addition, we aimed to understand the factors that recent graduates found valuable as they transitioned into the workforce. The final aim was to understand the amount of paid and unpaid work an individual typically completes before entering the sport science workforce.
2. Methods
2.1.
Participants
Full details of the methods used in the study including the target population, sample and recruitment strategies, instrument development,andanalyticmethodshavebeenreportedpreviously (Bruce et al., 2022); consequently, this information is briefly summarised in the following subsections.
A total of 116 participants completed the survey. Participants were recruited from the population of sport scientists in Australia. A purposeful recruitment strategy was used to recruit participants wherebyinformationaboutthesurvey wasdistributedtomembers of relevant state and national sporting bodies via membership and mailing lists, as well as circulated through social media and personal networks. Inclusion criteria for study eligibility included being engaged in the field of sport science (i.e., working or volunteering in sport science practice) in Australia during the survey period and aged over 18 years.
2.2. Survey instrument
Thesurveyinstrumentwasdevelopedbytheresearchteamandwas based on an initial set of questions from a previous survey of the sport science workforce (Dawson et al., 2013; Dwyer et al., 2019). It included questions separated into nine sections, with this paper reporting on demographic information and results from the section focusing on ‘Careers in sport science’. The demographic section included items for capturing participant information (e.g., age, gender, location, education) including current employment (e.g., number of jobs, status, sector, length). The ‘Careers in sports science’ section included 9 questions. Participants reported on the importance of 10 factors (e.g., ‘sports science allowed me to work with athletes’, ‘sports science was a career that suited my abilities’) for their decision to pursue a career in sport science using a 6-point rating scale (1 = not important, 5 = very important, 6 = unsure). They also reported on whether they were currently in a paid sport science role (not an intern or honorarium; yes, no) and the position titlethatbestdescribedtheirfirstpaidrole(notinternorhonorarium) in the sport science industry from a list of titles (e.g., ‘academic sportscience’,‘highperformancemanager’,‘performanceanalyst’). Those who were currently employed, reported on the number of unpaidandnumberofpaidvolunteerorinternroles theyhadbefore obtaining their first paid role (0, 1, 2, 3, 4, 5 or more), and those who were not currently employed, reported on the number of unpaid and number of paid volunteer or intern roles they had up to this point in their career (0, 1, 2, 3, 4, 5 or more). In addition, participants also reported on whether they had been working in the sport science industry for five years or less (yes, no); and for those reportinginvolvementforfiveyearsorless,theirlevelofagreement for 20 factors (e.g., ‘on the job training’, ‘internship/traineeship’, ‘professional networks’) that were helpful for their early work in the industry (1 = strongly disagree, 5 = strongly agree).
2.3. Procedure
Ethical approval for this study was obtained from the Deakin University Human Research Ethics Committee and all participants provided informed consent prior to completing the questionnaire. SurveydatawerecollectedusingtheREDCap(ResearchElectronic
DataCapture)software(Harrisetal.,2019).Datawascapturedover a 7-week period between October and December 2019 and survey completion took approximately 20 minutes.
2.4. Statistical analysis
Categorical demographic and career variables were summarised as proportions. Continuous variables were assessed for normality using published thresholds (Field, 2013; Lumley et al., 2002) and summarisedasthemean(andstandarddeviation).Forthis,‘unsure’ responses for the decision to pursue sport science career items were set as ‘missing’ and excluded from analysis. Ordered categorical variables were summarised as the median (and interquartile range [IQR]). All analyses were performedusing Stata16SE (StataCorp).
3. Results
Sample characteristics have been reported previously (Bruce et al., 2022). In brief, 116 participants were analysed (38 female, 78 male; < 25 years, n = 20; 26–35 years, n = 47; 36–45 years, n = 29; > 45 years, n = 21). Most participants were based in Victoria, New South Wales, and Queensland (total 81.2%); male (67.2%); aged 35 years or younger (57.8%); and hold a Master’s or PhD as their highest completededucation(67.5%),mostlywithinthefieldofsportscience (94%). Experience in the sport science workforce varied with the highest proportion of participants having less than 5 years’ experience (37.6%). This was followed by experienced practitioners with greater than15years’experience(29.9%),practitioners with6–9 years’ experience (20.5%), and practitioners with 10–15 years’ experience (12%).
Theimportanceofvariousfactorsforchoosingtopursueacareer in sport science are summarised in Table 1. Five factors (‘I was passionate about sport science’; ‘Sport science was a career that suited my abilities’; ‘I believed I would be a good sport scientist’; ‘Sport science allowed me to work with athletes’; ‘Sport science allowed me to work with coaches’) were rated moderately to very important (≥ 3.0) while the other factors were rated slightly to not important.
Table1:Importanceoffactorsforpursuingacareerinsportscience.
Factor
I was passionate about sports science
Sports science was a career that suited my abilities
I believed I would be a good sports scientist
Sports science allowed me to work with athletes
Mean (SD)
4.7 (0.6)
4.2 (1.0)
3.9 (1.1)
3.9 (1.2)
Sports science allowed me to work with coaches 3.4 (1.4)
Sports science provided me with opportunities to work overseas 2.6 (1.4)
As an athlete it was important for me to be good at sports science 2.4 (1.4)
I wanted to work in a sub-elite sport environment 2.2 (1.3)
Sports science provided opportunity for a high income 2.0 (1.2)
Sports science provided me with job security 1.9 (1.3)
Note: Number of responses varies across different factors; Rating scale: 1 = not important, 5 = very important
Atotalof82participantsreportedbeinginapaidsportscience role; 14 were not currently in a paid sport science role and data was missing for 21 participants. Those currently employed reported a median of 2.0 (IQR = 2.0) unpaid volunteer/intern and 1.0 (IQR = 2.0) paid volunteer/intern roles before obtaining their first paid sport science role. Whilst those not currently employed reported a median of 4.0 (IQR = 2.0) unpaid volunteer/intern and 1.0 (IQR = 2.0)paidvolunteer/internroles atthispointoftheircareer(i.e.,time of completing the survey). Those currently employed in a sport science role, nominated a range of position titles as best representing their first paid position (see Table 2). The most frequently nominated titles were ‘strength and conditioning coach’, ‘academic sport scientist’, and (generalist) ‘sports scientist’. Other common titles included ‘sports physiologist’, ‘performance analyst’, and ‘high performance manager’, while titles reflecting more specialised roles (e.g., ‘sports biochemist’, ‘sports biomechanist’, ‘sports dietitian’) were less frequently nominated. Almost two thirds (63%) of participants reported being in the same role currently as their first position.
Table 2: Position titles for first paid role in sport science.
Participants who had been involved in the sport science industry as a sport science professional for five years or less (n = 44, 37.6%) reported on factors that were most helpful to them in their early work in the industry. There was agreement (≥ 4.0) about helpfulness for a total of 12 separate factors with strongest agreement for two of these factors being ‘on the job training’ and ‘internship/traineeship’ (see Figure 1).
4. Discussion
Thisworksoughttounderstandwhypeopleenterthesportscience workforce, the factors that recent graduates find valuable as they transition into the workforce and the level of unpaid and paid (i.e., honorarium) work an individual needs to invest in prior to gaining paid employment. Participants pursued a career in sport science as they were passionate about the area, and it aligned with their perception of their abilities. People who were currently in a paid position reported having two unpaid and one paid position prior
Professional memberships
Experience working with non-healthy/para athletes
Networking with academic staff
Experience working with senior athletes
Formal mentoring arrangement
Professional development
Experience working with male athletes
Experience working with female athletes
Experience working with able bodied athletes
Experience working with junior athletes
Experience working in multiple sports
Experience working with sub elite athletes
Degree/qualification
Networking through practicum opportunities
Informal mentoring
Experience working with elite athletes
Professional networks
Experience working with variety of athletes
Internship/traineeship
On-the-job training
Agreement (1 = strongly disagree, 5 = strongly agree)
Figure 1: Importance of factors for early work in the sport science industry.
to their full-time employment, whilst those who were not in fulltime work reported having four unpaid positions and one paid position to date. Several factors were highly rated for assisting sport scientists in their current role with on-the-job training and internship/traineeship the most highly rated of these.
Participants pursued a career in sport science as they were passionate about the area, thought it suited their abilities and would allow them to work with athletes. These findings are consistent with previous work showing that people enter sport careers due to a love of sport and wanting to be involved in sport as a career (Mensch & Mitchell, 2008; Spittle et al., 2021; Vaartstra et al., 2017). Previously and concerningly (for sport scientists),in astudybyStevensetal.(2021)coachesrankedsport scientists (defined as specialists in the application of scientific principles and techniques to assist coaches and athletes improve their performance at an individual level or within the context of a team environment) as the lowest of eight practitioners (the other seven practitioners specified in the Steven’s et al. study were coach, physiotherapist, sports psychologist, high performance manager, strength and conditioning, sports doctor, dietician) around providing value for the athlete. This is despite agreement that sport scientists play a necessary role in sport and that they are effective in improving an athlete’s performance. This suggests that there may be a misalignment between expectations of sport scientists entering the workforce and the reality of the job
alongside the expectation of colleagues (e.g., coaches, physiotherapists; York, 2014). These findings indicate a need for industry to provide greater awareness of the roles and expectation required as a sport scientist.
Sport science, particularly at the high-performance level, has been considered a tough field to break into, with anecdotal stories of students quitting or changing career aspirations before gaining paid employment. Our findings revealed that those who were not currently working in a paid industry role had a higher number of paid and/or unpaid volunteer or intern roles compared to their peers who were working in industry. Participants who were not currently working in industry were mostly 1–3 years into their career (71%) with one 6–7 years into their career and another greater than 15 years. This indicates that due to unknown factors a considerable proportion may have struggled to break into the paid sport science role or are taking extra time to find their preferred role. Further research is required to understand these reasons.
Within exercise and sport science, 41% of graduates had to volunteer before being hired and paid in the workforce (Stevens et al., 2021). Our findings show even higher levels of volunteer work prior to obtaining a first paid role in sport science with 89% of participants reporting having at least one unpaid position and 65% at least one paid volunteer or intern role. It appears that for sport scientists, there remains an expectation by employers to
obtain additional experience before gaining a paid role. This highlights the potentially challenging nature of sport science and suggests it may be harder to specialise in than other exercise science career options (e.g., accredited exercise physiologist, exercise scientist, physiotherapy) given the larger percentage of graduates who have undertaken unpaid work before gaining employment compared to the Stevens et al. (2021) findings.
The necessity to complete intern or volunteer roles was however identified as one of the most important factors for early work in the industry alongside ‘on the job training’. Students graduating from ESSA accredited courses are required to undertake WIL which is designed to provide students with an opportunity “to develop and demonstrate competence in integrating and applying their professional knowledge and skills in a real-world setting” (Exercise and Sports Science Australia, 2022b, p.3). It seems that recent graduates entering the sport science workforce need more experience than is provided during their WIL as evidenced by the internships completed postgraduation. Students completing an exercise and sport science undergraduate degree are not required to complete WIL within sport science (rather Exercise Science), so this may indicate that they are then seeking additional sport science experiences to gain the required skills and knowledge for the sport science industry. Further research is required to understand what is contributing to the current situation; for example, but not limited to, is WIL providing the necessary opportunities for students to experience sport science? What is the responsibility of professional colleagues in industry alongside ESSA to work towards a solution? And, what are the ethical considerations of graduates completing multiple unpaid internships? among other questions to be considered.
Internships are often longer than WIL experiences (advertised as 10–12-month opportunities) and as a result interns may feel more ‘a part’ of the organisation than when completing WIL, thus providing a more authentic experience. Whilst this may be of benefit to a graduate student, it is likely only those with underlying financial security can take on these internships or multiple unpaid roles and absorb the loss of salary, creating inequity within the field. Further research should examine what it is about on the job training and internships that recent graduates find valuable, with the intent to embed this, where possible, into universitydegrees. Understanding the needs of both theuniversity sector and industry will be challenging, alongside considering the ethical implications of additional WIL or internships, especially those which are unpaid.
Job security and a high income were not considered to be the most important factors for pursuing a sport science career. These commonly cited characteristics of sport science may act as barriers to attracting people into the industry. Low job security has been observed as a potential issue by Stevens et al. (2021) who found that sport scientists in sporting clubs are likely to be in their position for less than five years. Alongside job security, ‘(lack of) demand’ has been identified as a low scoring reason for pursuing a career in sport science (Spittle et al., 2021), thus are areas for concern of potential sport scientists. Previously, sport scientists within sports teams ‘somewhat disagreed’ with the statement that ‘sport scientists generally receive fair working conditions’ (Stevens et al., 2021). There is an opportunity for an increase in education and awareness of the issues faced by sport scientists and how to be valued in the workforce.
This body of work focused on understanding sport scientist’s perceptions of their early career trajectory and the factors contributing to success in their initial working career. The sample did not reach all sport scientists employed in Australia and may represent the views of only a sample of sport scientists. Sampling may have been biased towards those working in sport science and may not have captured those who are still seeking roles within the industry or who have already existed the field due to reasons that may include lack of opportunities, limited career pathways or low levels of satisfaction. Future research may look to understand the career paths of those who pursue a different career due to the lack of opportunities within high performance (elite) sport.
Overall, findings of this study supported previous research showing that people pursue a career in sport science in line with their passion and to align with their perceived abilities. They also revealed that sport scientists are required to gain more experience (e.g., internship) than may be presented to them by WIL opportunities. This may be due to undergraduate degrees requiring WIL within the realm of Exercise Science, and WIL for sport science only being required if they complete a postgraduate degree. Further research is required to understand the benefits and practicalities of longer internships alongside the potential issue of inequity. This is particularly important as these experiences are perceived as valuable once employed in industry as ‘on the job training’ and ‘internship/traineeships’ were the highest rated experiences for assisting them in their current role.
Conflict of Interest
The authors declare no conflict of interests.
Acknowledgment
The authors acknowledge Caleb Lewis and Tom Eaton who assisted with the survey development and recruitment for the study. We would also like to thank the participants who provided their time to complete the survey. Funding to complete the project was received from the School of Exercise and Nutrition Sciences, Deakin University.
References
Bowes, A., & Culvin, A. (2021). Introduction: Issues and debates in the professionalisation of women’s sport. In A. Bowes & A. Culvin (Eds.), The professionalisation of women’s sport (pp. 1–15). Emerald Publishing Limited. https://doi.org/10.1108/978-1-80043-196-620211001
Bruce, L., Bellesini, K., Aisbett, B., Drinkwater, E. J., & Kremer, P. (2022). A profile of the skills, attributes, development, and employment opportunities for sport scientists in Australia. Journal of Science and Medicine in Sport, 25(5), 419–424. https://doi.org/10.1016/j.jsams.2021.12.009
Dawson, A., Wehner, K., Gastin, P., Dwyer, D. B., Kremer, P., & Allan, M. (2013). Profiling the Australian high performance and sports science workforce. Deakin University. http://hdl.handle.net/10536/DRO/DU:30060700
Dwyer, D. B., Bellesini, K., Gastin, P., Kremer, P., & Dawson, A. (2019). The Australian high performance and sport science workforce: A national profile. Journal of Science and
Bruce et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 14-19 (2024)
Medicine in Sport, 22(2), 227–231. https://doi.org/10.1016/j.jsams.2018.07.017
Exercise and Sports Science Australia. (2008). Annual Report 2008. Exercise and Sport Science Australia. https://www.essa.org.au/common/Uploaded%20files/Reports /annual-report-2008.pdf
Exercise and Sports Science Australia. (2019). Sports science graduate internship guidelines Exercise and Sports Science Australia. https://www.essa.org.au/Public/Public/Advocacy/Policy_Stat ements.aspx
ExerciseandSportsScienceAustralia.(2021). Annual Report ’21 Exercise and Sport Science Australia. https://www.essa.org.au/common/Uploaded%20files/Reports /annual-report-2021.pdf
Exercise and Sports Science Australia. (2022a). Courses with full or provisional accreditation status Exercise and Sport Science Australia. Retrieved from https://www.essa.org.au/Public/EDUCATION_PROVIDERS /Courses_with_Full_or_Provisional_Accreditation_Status.as px
Exercise and Sports Science Australia. (2022b). ESSA Practicum Standards Exercise and Sport Science Australia. Retrieved from https://www.essa.org.au/Public/EDUCATION_PROVIDERS /Applying_for_Full_Accreditation.aspx
Field, A. (2013). Discovering statistics using IBM SPSS statistics (4th ed.). SAGE.
French, D., & Torres-Ronda, L. (2022). NSCA’s essentials of sport science. Human Kinetics.
Harris, P. A., Taylor, R., Minor, B. L., Elliott, V., Fernandez, M., O’Neal, L., McLeod, L., Delacqua, G., Delacqua, F., Kirby, J., & Duda, S. N. (2019). The REDCap consortium: Building an international community of software platform partners. Journal of Biomedical Informatics, 95, 1–10 https://doi.org/https://doi.org/10.1016/j.jbi.2019.103208
Lumley, T., Diehr, P., Emerson, S., & Chen, L. (2002). The importance of the normality assumption on large public health data sets. Annual Review of Public Health, 23, 151–160. McLachlan, F. (2019). It’s boom time! (Again): Progress narratives and women’s sport in Australia. Journal of Australian Studies, 43(1), 7–21. https://doi.org/10.1080/14443058.2019.1575262
Mensch, J., & Mitchell, M. (2008). Choosing a career in athletic training: Exploring the perceptions of potential recruits. Journal of Athletic Training, 43(1), 70–79. https://doi.org/10.4085/1062-6050-43.1.70
Spittle, M., Daley, E. G., & Gastin, P. B. (2021). Reasons for choosing an exercise and sport science degree: Attractors to exercise and sport science. Journal of Hospitality, Leisure, Sport & Tourism Education, 29, 1–10 https://doi.org/10.1016/J.JHLSTE.2021.100330
Stevens, C. J., Lawrence, A., Pluss, M. A., & Nancarrow, S. (2018).The careerdestination,progression, andsatisfaction of Exercise and Sports Science graduates in Australia. Journalof Clinical Exercise Physiology, 7(4), 76–81. https://doi.org/https://doi.org/10.31189/2165-6193-7.4.76
Stevens, C. J., McConnell, J., Lawrence, A., Bennett, K., & Swann, C. (2021). Perceptions of the role, value and barriers of sports scientists in Australia among practitioners, employers and coaches. Journal of Sport and Exercise Science, 5(4), 285–301. https://doi.org/10.36905/jses.2021.04.07
Vaartstra, M. B., Kercher, V. M., Start, A., Brown, A. N., Peterson, M. D., & McGrath, R. (2017). Understanding why undergraduate students declare and continue to study an exercise science-related major. International Journal of Exercise Science, 10(5), 807–817.
York, R., Gastin, P., & Dawson, A. (2014). What about us? We have careers too! The career experiences of Australian sport scientists. International Journal of Sports Science & Coaching, 9(6), 1437–1456.
The Journal of Sport and Exercise Science, Vol. 8, Issue 1, 20-32 (2024) www.jses.net
Pattern recognition in soccer: Perceptions of skilled defenders and experienced coaches
James Feist1* , Oliver R. Runswick2 , Ed Hope3 , Jamie S. North4, Chris Pocock1
1University of Chichester, Institute of Applied Sciences, United Kingdom
2King’s College London, Department of Psychology, Institute of Psychiatry, Psychology & Neuroscience, United Kingdom
3Liverpool John Moores, Faculty of Science, School of Sport and Exercise Sciences, United Kingdom
4St. Mary’s University, Twickenham, Faculty of Sport, Allied Health and Performance Science, United Kingdom
A R T I C L E I N F O
Received: 13.02.2024
Accepted: 01.07.2024
Online: 23.09.2024
Keywords:
Perceptual-cognitive skill
Competition
Game reading
Practice environments
Coaches
A B S T R A C T
The ability to perceive and recognise patterns of play is important for performance in tasks with strict spatiotemporal constraints. Study aims were twofold: (i) to qualitatively investigate the mechanisms and processes underpinning how soccer players recognise patterns, (ii) to qualitatively investigate the importance of pattern recognition in competition and practice environments. Six skilled soccer central defenders and seven experienced soccer coaches were interviewed. A reflexive thematic analysis of the data identified six higher-order and twenty-two lower-order themes relating to pattern recognition and anticipation in competition and practice environments. The six higher order themes were: recognising danger and distance to ball, sources of information, experience, opposition team, organisation and communication, and development in practice environments. Participants shared that developing pattern recognition and game reading skill is crucial in creating effective practice environments that support the transfer of skills into competition. Providing central defenders with representative scenarios during practice is recommended to stimulate problem-solving and promote familiarity with patterns of play to underpin game reading and thus skilled performance
1. Introduction
Perceptual-cognitive expertise is frequently described as the ability of an individual to process information from the environment and integrate with existing cognitive knowledge structures to produce an appropriate response (Marteniuk, 1976). A contrasting approach pertains to a more direct perception viewpoint on expertise (Gibson, 1979). Specifically, from this perspective, individuals adapt their movements to the interacting constraints of a performance environment by continuously perceiving information to regulate goal-directed actions and govern skilled behaviour (Seifert et al., 2013). Regardless of the theoretical lens adopted, researchers have long been interested in studying the perceptual-cognitive skills of expert performers to identify the underpinning mechanisms that contribute to their expertise (e.g., Abernethy & Russell, 1987; Hodges et al., 2021; Ward & Williams, 2003). The importance of perceptual-cognitive
skills is magnified in dynamic and temporally constrained tasks such as medicine (Bertram et al., 2013), aviation (Russo et al., 2005) and sport (North et al., 2016; Roca et al., 2013), where performers must not only select appropriate responses, but do so under strict time pressure in the context of an ever-changing environment.
In this study, we adopt an indirect perception (i.e., information processing) approach to expert performance, highlighting the important interaction between cognition and action, as per the previous definition provided by Marteniuk (1976). In sports such as soccer, perceptual-cognitive expertise is demonstrated through the ability of skilled performers to think ahead and anticipate the actions of opposing players. In the present study we conceptualise anticipation in soccer through the term ‘game reading’, with this term used frequently by soccer coaches as an interchangeable term for anticipation (Williams & Jackson, 2019). Den Hartigh et al. (2018) investigated ‘game reading’ in youth soccer and found
*Corresponding Author: James Feist, Institute of Applied Sciences, University of Chichester, United Kingdom, J.Feist@chi.ac.uk
that players who were selected for a soccer school of a professional club were capable of processing and structuring game-related information tohigher levels of cognitive complexity (Den Hartigh et al., 2018), when compared to their non-selected counterparts. Moreover, the selected players were better able to identify structured information on the field (e.g., positioning of teammates) and integrate relational information between players (Den Hartigh et al., 2018). Therefore, gaining an insight into how soccer players recognise patterns of play would provide further depth to our understanding of anticipation (‘game reading’) in soccer.
The ability to perceive and recognise patterns between features (i.e., players; Abernethy et al., 2005; North & Williams, 2019) is a perceptual-cognitive skill that has been proposed as important in ‘reading the game’ (cf. Williams et al., 2011). Following extended domain-specific practice, highly skilled performers are proposed to develop specialised cognitive knowledge structures which enable them to attend to the most important sources of information and disregard non- or lessrelevant information (Ericsson & Kintsch, 1995; Ericsson et al., 1993; 2009). Together, this enables skilled performers to reduce the complexity of a display and encode information more efficiently. Information can then be evaluated and considered against previously encountered situations (North et al., 2011). Thus, experts are able to retrieve relevant task-specific information from their long-term working memory (Ericsson & Kintsch, 1995), which enables performers to anticipate quickly and accurately through the process of perceiving and recognising patterns.
An extensive body of laboratory-based experimental research has been conducted to extend our knowledge of pattern recognition (e.g., Allard et al., 1980), as well as understanding its underlying mechanisms (e.g., Williams et al., 2012). This includes attempts to understand the nature of information that underpins pattern recognition expertise employing techniques such as spatial occlusion methods (e.g., Williams et al., 2006), eye-movement data (e.g., van Maarseveen et al., 2018) and verbal protocol analysis (e.g., North et al., 2011). This body of research has provided evidence as to the importance of relative motion information between key players for pattern recognition. More specifically, when viewing offensive sequences of play, expert defenders rely on relations between centrally located attacking players as well as the player in possession of the ball to guidetheir familiarity judgements (e.g., North et al., 2017). However, there has been a notable lack of research that investigates how players and coaches perceive the importance, underpinning processes, and development of pattern recognition.
Researchers have presented evidence to understand the processes and mechanisms underpinning recognition and anticipation in soccer (North et al., 2009; North et al., 2011) and the relative importance of different perceptual-cognitive skills during anticipation (North et al., 2016). Using eye-movement analysis and verbal reports, North and colleagues (2009; 2011), found performance of skilled players on respective anticipation and recognition tasks to be positively correlated to a moderate extent. North et al. (2016) employed an anticipation paradigm in which skilled and less-skilled soccer players made anticipation judgments to stimuli in ‘near’ (ball in close proximity) and ‘far’ (ball far away) conditions that were presented in both film and point light display format to manipulate the availability of
information from postural cues. When the ball was far away anticipation performance was unaffected by presentation mode, suggesting an important contribution of pattern recognition to anticipation. This quantitative research provides evidence as to link between pattern recognition and anticipation.
Qualitative investigations havethepotential to provide adepth and richness of information that is not possible through traditional experimental protocols (Greenwood et al., 2014). This method of data collection provides an opportunity to ask skilled and experienced performers and coaches directly to better understand the construct under investigation. In recent years there has been a growth in researchers making use of qualitative research methods to investigate topic areas in skill acquisition and expert performance. For example, researchers have employed qualitative methods to understand practice design activities for visual exploratory activity in soccer (Eldridge et al., 2023), and factors affecting soccer coaching behaviour (Jewell et al., 2022). However, there remains a lack of work aimed at understanding player and coach perceptions of pattern recognition.
While not in the domain of soccer specifically, Carboch et al. (2023) conducted nine interviews with active or former professional tennis players (current coaches) exploring how players use anticipatory information sources for serving and returning inelite tennis.Afterconducting an open-coding analysis approach, the authors identified pattern recognition as a lower order theme. Participants reported that pattern recognition varies depending upon skill level and the authors highlighted the importance of players gathering information from various sources both before and during matches to support anticipation (Carboch etal.,2023),suchascontextualinformation(Murphyetal.,2018). A rare example of using qualitative methods to investigate perceptual skill in soccer utilised interviews with professional players to understand contextual factors influencing decisionmaking (Levi & Jackson, 2018). Following an inductive thematic analysis, there were four dynamic (e.g., situational development of the match) and three static (e.g., external context of the match) themes that influence the ability of skilled soccer players to make confident, effective decisions during a match. From the perspective of skilled players, the ability to recognise and practice familiarpatternsofupcoming opponentsinrepresentativetraining scenarios can help to support decision-making in competition (Levi & Jackson, 2018). Participants in the Levi and Jackson (2018) study also discussed how contextual factors (e.g., score, momentum, perceptions of performance) can influence the decisions made in competitive soccer matches. However, whilst these findings provide an initial and valuable insight to factors which affect decision-making, they do not offer in-depth analysis of how soccer players recognise patterns of play in competition, despite its apparent importance to expertise in soccer (Abernethy et al., 2005).
There is a paucity of research investigating what the players and coaches understand to be vital with regards to 'reading the game' and recognising patterns of play. This is somewhat surprising considering the volume of quantitative data amassed from previous experimental approaches. Moreover, and perhaps indicative of the general research area, there is a lack of research that has sought to garner the experiential knowledge of coaches, and not only their perceptions of what constitutes pattern recognition, buthow they seek todevelop this skill withinpractice activities. Therefore, our study aims were twofold: (i) to
qualitatively investigate the mechanisms and processes underpinning how soccer players recognise patterns, and (ii) to qualitatively investigate the importance of pattern recognition in competition and practice environments.
2. Methods
2.1. Philosophical assumptions
To understand the participants’ perceptions of the importance of pattern recognition to ‘game reading’ in soccer, we used a qualitative approach that was underpinned by interpretivism and a relativist ontology. Relativism accepts that multiple subjective, butequallyvalid,realitiesoftheworldcanexist(Sparkes&Smith, 2014). The approach used was framed epistemologically by constructionism, which considers knowledge as subjective and socially constructed (Smith & McGannon, 2018).
2.2. Participants
Six skilled soccer central defenders (Mage = 23 years, SD = 5.5; Mplaying experience = 14.1 years, SD = 4.1) and seven experienced soccercoaches(Mage =43years,SD=10.5; Mcoaching experience =23.1 years, SD = 3.6) were interviewed. Skilled central defenders were required to satisfy two criteria: to play semi-professional soccer or higher in the previous five years and classify themselves as a central defender (see Table 1 for details). Central defenders were recruited due to pattern recognition research in soccer typically
asking players when viewing film-based stimuli to imagine themselves as a central defender (e.g., North et al., 2017) as this position allows defenders to see the whole game in front of them. The coaches were required to satisfy two main criteria: to possess a UEFA B qualification in football (soccer) coaching or higher and to have a minimum of 10 years of soccer coaching experience (see Table 2 for details). Using a criterion-based purposive sampling of defenders and coaches ensured that participants had appropriate experiences to discuss for the study, and the sample size offered sufficient experiences to address the aim of the study (Malterud et al., 2016). Ethical approval was granted from the lead author’s institution and all participants provided written informed consent.
2.3. Interview guide
Semi-structured interviews were used to allow the emergence of unforeseen topics, with participants encouraged to share their knowledge and talk about personal experiences (see Smith & Sparkes, 2016). The interview guide comprised of five sections: soccer background, understanding of pattern recognition/game reading in soccer, use of pattern recognition/game reading in competition environments, pattern recognition/game reading in practice environments and the overall importance of pattern recognition/game reading in soccer. The lead author used commonly used coaching terms such as ‘game reading’ when conducting interviews with all participants.
Table 1: Participant characteristics of the six soccer central defenders that were interviewed
Player Gender Age (y) Experience (y) Highest level played Current level of play
P1 Male 19 12 UK Professional Men’s Academy UK Men’s Semi-Professional
P2 Female 26 16 International Women’s Soccer UK Professional Women’s Soccer
P3 Male 18 10 UK Professional Men’s Academy US College Soccer
P4 Female 32 20 International Women’s Soccer UK Professional Women’s Soccer
P5 Male 25 17 UK Professional Men’s Academy UK Men’s Semi-Professional
P6 Female 19 10 UK Professional Women’s Soccer UK Professional Women’s Soccer
Notes: ‘Highest level played’ refers to the highest standard a player has competed. For example, if a player’s highest competitive level of soccer played is ‘International’, this player has represented their national team in a competitive international match.
Table 2: Participant characteristics of the seven male soccer coaches that were interviewed
Coach Age (y) Experience (y) Highest level coached Current level of coaching Highest qualification
C1 38 21 UK Professional Men’s Academy UK Independent Men’s Academy UEFA A License
C2 49 31 UK Men’s Semi-Professional UK Men’s Semi-Professional UEFA A License
C3 40 23 UK Professional Women’s Soccer
C4 38 23 UK Men’s Semi-Professional UK Men’s Semi-Professional UEFA B License
C5 36 20 UK Professional Men’s Academy UK Professional Men’s Academy UEFA A License
C6 65 22 UK Professional Men’s Academy UK Men’s Semi-Professional UEFA B License
C7 36 22 UK Professional Women’s Soccer UK Independent Men’s Academy UEFA A License
The interview guide was developed through: (i) quantitative findings in pattern recognition in soccer; (ii) the lead author’s experiences of playing and coaching soccer (18 years playing experience and 8 years coaching experience); and (iii) assessing previous qualitative studies’ interview guide approaches (e.g., Morris-Binelli et al., 2020; Pocock et al., 2020). Example questions posed to both soccer players and coaches were “are there any similar patterns of play in soccer that you see occur all the time?” and “in your own words, how important do you think pattern recognition is for a central defender in soccer?” The full interview guides used in this study are available as a supplementaryfile.Apilotinterviewwasconducted withacentral defender with 15 years of competitive playing experience. The interview was then reflected upon, with minor changes made, includingremovalofcertainquestionsduetoperceivedrepetition.
2.4. Data collection
All interviews were conducted by the lead author on an individual basis via Zoom (version: 5.4.9. 59931.0110). Participants were provided with a written information sheet, as well as a verbal description of the study before the interview. Furthermore, participants were made aware of the interview being recorded via an MP3 device as well as the confidentiality of their responses and their right to withdraw. Open-ended questions allowed for detailed descriptionsofexperiences(Smith&Caddick,2012)andwereused to build a discussion around the players’ and coaches’ knowledge and understanding of pattern recognition and game reading. Probe questions were used to encourage further articulation of points and build rapport (Smith & Caddick, 2012; Smith & Sparkes, 2016). Interviews ranged between 39 and 54 minutes (M = 46 minutes, SD =5.3)andwereaudio-recordedonanMP3devicefordataanalysis.
2.5. Data analysis
Interviews were transcribed verbatim with grammatical changes made where necessary. To ensure anonymity, each participant was assigned an identifying label (i.e., coaches were labelled C1 to C7, and players labelled P1 to P6) to identify similarities and differencesmoreeasily(seeMorris-Binellietal.,2020).Areflexive thematic analysis approach was adopted, which allowed for identifying themes derived from the raw data (Braun & Clarke, 2019).Thisreflexiveapproachwascentredaroundtheleadauthor’s role in knowledge production as well as the transparency and consistency of analytical decisions made during the analysis phase (Braun & Clarke, 2019). Towards the end of the coding process, a more collaborative approach was adopted to achieve greater depth when understanding the data. As part of this approach, the current investigation did not exclusively use either a deductive (structured, theory-driven approach) or inductive (little use of theory, framework, or pre-determined structure) approach. Instead, a pragmatic form of enquiry was selected, which applied both approaches (see Braun et al., 2016; Strafford et al., 2021). Understanding of previous literature related to pattern recognition in soccer, as well as the lead author’s experiences of playing and coaching soccer, informed this pragmatic approach. To acknowledge and minimise any preconceived biases, the lead author took steps to explore alternative viewpoints during the interpretation stage of the analysis process. This involved discussing all codes and ideas with all co-authors. The current
investigation followed the six phases of reflexive thematic analysis detailedbyBraunandClarke(2006;2019).Aspartofthisapproach, a recursive rather than linear process was utilised for data analysis, as the lead author actively engaged with and immersed themselves in the data (Morris-Binelli et al., 2020).
The first stage of the analysis process followed a deductive approach where the lead author read each interview transcript multiple times to specifically identify language relating to pattern recognition and ‘game reading’. Following this deductive approach, a strictly inductive approach was adopted to identify key themes from the raw data (Braun & Clarke, 2006). The third phase consisted of coding and collating units of text which were then reorganised and used to identify potential themes from the codes. Following this, the fourth phase involved initial themes being refined and sub-themes being formed which led to the creation of the ‘thematic map’. In the fifth phase, the lead author conducted regular, meaningful reflections on the proposed themes. At this stage, the addition of a ‘critical friend’ (a lecturer in sport and exercise psychology) allowed for continual discussion and reflection on each theme and sub-theme to ensure the themes accurately reflected the data. Any coding differences that were identified were resolved through discussion and codes were altered if deemed appropriate (Strafford et al., 2021). The final phase involved writing up the analysis with the most relevant quotes from participants beingselected to capture each high-order theme presented.
2.6. Methodological rigour
Two main processes were employed to ensure methodological rigour. First, four participants were randomly selected for member reflections (Tracy, 2010) which involved sending completed transcripts and an overview of the results to participants (Smith & McGannon,2018).Nochanges weremadeto the transcripts or data analysis as a result of the member reflections. Nowell et al. (2017) emphasises how member checking can be used to enhance the credibility of the analysis process. Second, to enhance the confirmability of the research, an independent critical friend discussed the lead author’s interpretations of the analysis. The critical friend was valuable during data analysis, providing constructive comments which enabled the lead author to reflect and defend their judgements concerning the proposed themes that were constructedfromthedata(Burke,2016;Smith&McGannon,2018).
3. Results
Prior to understanding perceptions of skilled defenders and experienced coaches on the importance of pattern recognition to anticipation (‘game reading’) in soccer competition and its developmentinpracticeenvironments,alogicalstartingpointwas to ask all participants to provide their definitions of the term ‘game reading ’ Coach 1 described ‘game reading’ as “thinking into the future so identifying where the space is, where opposition players are, team members, where the ball is on the pitch.” Similarly, Player 5 defined ‘game reading’ as “being able to adjust your position or your thought process to what you’re seeing in front of you,” with Player 3 describing ‘game reading’ as “reading the person, reading the situation and anticipating what’s going to happen next. ”
Thedefinitionsprovidedbycentraldefendersandcoachesshare a number of similarities. All quotes point to a player’s ability to: (i) recogniseemergingpatternsof playbeforetheyunfold; (ii) identify relevant kinematic cues from opposition players; and (iii) continually adjust one’s body position whilst visually exploring the environment to locate teammates, the ball, opposition players and empty space. In summary, based upon the perceptions of participants in this study, the key contributing factors to ‘game reading’ (i.e., anticipation) were kinematic cues, situational probabilities, visual information, and pattern recognition. While
participants highlighted a number of perceptual-cognitive skills underpinning game reading in soccer, in the context of the present study,wewerespecificallyinterestedingarneringtheirperceptions oftheimportanceofpatternrecognitiontothisprocess.Tothis end, a two-stage reflexive thematic analysis of the data resulted in the generation of twenty-two lower order themes and six higher-order themes (see Figure 1). The six higher-order themes are discussed and are supported with illustrative quotes from central defenders and coaches.
perspectives ofskilledsoccer central defenders and experienced coaches.
Figure 1:Thematicmap ofpattern recognition and anticipation incompetition andpracticefrom the
3.1. Recognising danger and distance to ball
Both players and coaches referred to the importance of central defenders recognising dangerous situations and their distance to the ball which appears an important part of pattern recognition. Central defenders are required to frequently move their body and head to visually explore (‘scan’) their surrounding environment when out of possession to identify the movements of teammates and opposition players. Players and coaches also described how central defenders are required to have excellent ‘positional awareness’ and an ability to ‘recognise triggers’ (e.g., off the ball movements of central attacking players) which appeared heavily linked to ‘ball location’ and ‘pressure on the ball’. Player 5 describes:
“I'd want to be in a position so I can see the ball and the man I am marking. I want to be able to see both of those, and then across that I'm probably looking at my other centre back and my holding midfield player. Are they matched up in a good position of where the ball is going to be because even though the ball is quite far away from us it can travel quickly… as you get closer to the ball, then you become more interested in the ball rather than the runners and positioning around you” (P5).
Players explained how, when recognising emerging patterns of play, the ball's location combined with their positional understanding (identifying where they are standing on the pitch in relation to team-mates and opposition players), had a direct impact upon the prevention of dangerous situations. This is consistent with previous literature that identified how the position of the ball in relation to opposition players and the information that arises as a result of the interactive movements of team-mates, are critical structural information, linked to anticipation and recognition ability (see also Roca et al., 2013; North et al., 2016). Coach 7 explains:
“… their [central defenders] ability to move forwards,backwards,sideways,justas theballdoes, to get into positions, quickly… I think it comes down to those to understand if there's a trigger or pattern happening, but they've got to see that early” (C7).
Taken together, these findings suggest that during competitive performance,acentraldefenderisrequired tocontinuallyevaluate their position whilst simultaneously identifying out of possession triggers to facilitate the recognition of structured patterns of play. Player 4 shared:
“I'm looking at how deep the back line is. I'm lookingatwheretheballis,I'mlookingatourshape. Are we narrow and compact? Are we cutting channel balls out, are we checking shoulders and picking up runners?” (P4).
This highlights how central defenders are visually scanning to pick out the most information-rich areas by checking their shoulders and processing each individual moment as the game
unfolds in front of them. Research has detailed how players who visually explore their environments more effectively can perceive approaching opponents, identify teammates, and execute more appropriate passing options (e.g., Jordet et al., 2020). Player 4 further recalled an in-game situation where they recognised their active role in influencing the pattern of play:
“If that midfielder hadn't been dragged out for us, we knew that they probably weren't going to try and break that line pass. So, they might have gone wide. So that's when the fullbackknew that was their time, that was their trigger, and I would cover round, and the midfield would just kind of screen” (P4).
3.2. Sources of information
Both players and coaches highlighted the importance of a central defenders’ ability to utilise various sources of information from the environment. More specifically, participants explained how central defenders draw upon postural cues of opposition players to predict likely upcoming situations. Players and coaches regularly acknowledged the importance of the ‘relative motion information between central attacking players’ be fundamental to successful pattern recognition and ‘game reading’. Player 1 shared:
“… the minute I saw that ball go there I'll see his head come up and his foot go back, he's going to play the ball… so, as I said, it was all about that, in each split second of movement, and where they were in analysing their striker’s movement, or their winger’s movement, it's just that really quick decision-making recognizing that pattern of play and where they're going to ultimately try to exploit” (P1).
This suggests there are various aspects of situational information that interact in a single phase of play. It appears that skilled central defenders can use their enhanced knowledge of game-specific event probabilities (Williams et al., 2005) and assign an abstraction hierarchy to select and recognise which information source is most pertinent in a given situation to inform their anticipatory judgements. Additionally, this links to the finding that when the ball is further away, the perception of structured patterns is most important, whereas when the ball is closer to thegoal,recognitionof an opponent’s bodily movements provide the most important information for successful anticipation (Roca et al., 2013; North et al., 2016). One coach reported:
“… looking at their centre forwards’ forward runs and runs in behind, identifying key players, key strengths, recognising movements that they make…” (C3).
The above quote from coach 3 pertaining to the movement of central attacking players supports previous literature which has provided evidence to suggest that micro-relations between central attacking players are vital for successful pattern recognition for
bothlocalandglobal(i.e.,wholefield)patterns(Hopeetal.,2024; North et al., 2017; Williams et al., 2012).
3.3. Opposing team
Central defenders and coaches gave accounts of how the opposition team’s formation, style of play, in-possession tendencies, and gestures fromopposition players link to a player’s ability to recognise patterns of play and ‘read the game’. To exemplify, Player 3 said:
“What are their tendencies? What are they looking to do? Are they looking to try and aim for the centre forward quite a lot? … So, I think the style of play and the patterns depends a lot on the team and where we are in the world … so that would depend on who you're playing” (P3).
Researchers have shown that soccer teams have behavioural patterns that shift throughout a competitive match where more common phases of play will occur, suggesting the importance of familiarity of the opposition team’s in-possession tendencies (Gonçalves et al., 2019). This experiential knowledge from Player 3 reveals that opposition teams perform similar movement patterns based upon their playing philosophy and tactical shape. Coach 1 states:
“I think it [pattern recognition] depends on who you're playing against. I think it depends on teams, you know, playing styles and the opposition that you're faced with and yeah, I mean, depends on team’sformation,setup,philosophy,belief…”(C1).
This additionalcontextual information,throughtheanalysis of future opposition and their preferred formation and style of play, may aid central defenders' ability to recognise structured sequences of play.
3.4. Organisation and communication
Players and coaches articulated how during competitive performance, a central defender’s role involves organising teammates into suitable positions and reacting to information from coaches. Participants explained the importance of central defenders using their voice to talk teammates through the game and intospecific pitchlocations to aid therecognitionof emerging patterns to support ‘game reading’. Player 5 explains:
“… if I can stop the ball going into the centre forward by placing a man in front by calling one of my midfield players in front of that, that's me doing my job and doing my job better than if the ball going into the centre forward, and then me having to make a tackle…” (P5).
Evidence has suggested that as central defenders are positioned centrally on the pitch, this allows them to share beneficial task-specific information with team-mates (McLean et al., 2021). This may imply that regular use of verbal
communication in the form of directing team-mates around the pitch as the ball is in transition may act as an anticipatory mechanism to prevent the ball from going into dangerous areas. Outside of soccer, previous literature has investigated the role of teammate communication in lacrosse (Riches et al., 2021) and beach volleyball (Klatt & Smeeton, 2020). Both studies provide evidence for the important role teammate communication plays in team sports to support decision making (Klatt & Smeeton, 2020) and anticipation (Riches et al., 2021). It is worth highlighting how a central defender’s ‘organisation and communication’ should be considered relevant to pattern recognition and ‘game reading’ skills by organising team-mates into relevant positions. However, there may central defenders who are skilled at ‘game reading’, yet may lack leadership and organisation skills to communicate effectively with team-mates. Coaches can help develop the communication skills of defenders, as well as providing instructions to develop game reading skills. Coach 5 describes instructions provided to a central defender:
“So, for example, I might say, ‘open your body up so you can see that the wide player and you can see the ball” (C5).
Therefore, central defenders use frequent communication to organise players around themand receiveverbal instructionsfrom coaches, especially regarding areas of perceived weaknesses (Levi & Jackson, 2018). This example of precise, positionspecific information provided to a central defender may be of value if a player struggles to recognise patterns of play. However, other than the work of Smith and Cushion (2006) who investigated the in-game behaviours of six professional soccer coaches and identified ‘developing game understanding’ as a key theme underpinning their behaviour, there is a lack of research conducted on identifying the in-game instructions provided to players by coaches.
3.5. Experience
Both coaches and central defenders articulated how experience of previous match and training situations appear to support central defenders in the recognition of frequently occurring patterns of play, the subconscious recognition of environmental information, the ability to problem solve and develop positional relationships with the defensive unit through repetition of game scenarios:
“I don't actively think about myself recognising a pattern… I think a lot of it is subconscious if that makes sense because from my point of view, I've had a lot of football [soccer] by the time I was 18 at Academy level,Ithink Iwas capable ofrecognising patterns in games …so I think it was always at the back of my mind that I felt confident that I knew what to do when certain patterns were unfolding be it striker coming deep and attacking midfielder running in behind me” (P2).
Despite suggestions that pattern recognition may be a byproduct of experience rather than a direct contributing characteristic to expert performance (North & Williams, 2019),
the quotes captured in this study indicate that skilled soccer central defenders consider previous experience an important aspect that directly contributes to their expertise. One coach explains:
“A lot of centre backs who are a bit older might not be as quick but the experience on how they understand the game, they're starting to rehearse every scenario of the game that's going on in front of them and they get themselves into that position very early” (C4).
These findings can be considered in relation to the theory of long-term working memory (Ericson & Kintsch, 1995) in which skilled performers develop domain-specific knowledge structures through extended practice, and can ‘chunk’ (Chase & Simon, 1973) meaningful information together, allowing them to rapidly retrieve information in the long-term working memory to guide anticipatory judgments (Roca et al., 2011).
3.6. Development in practice environments
All central defenders and coaches acknowledged the importance of practice to develop pattern recognition and game reading skills through match realistic opposed practises, unopposed practices, and coaching through the game:
“The game paced ones, the ones that were as realistic to games as possible, would benefit me the most… so, for me, it was doing stuff at match pace, whether it be even in a small-sided game or attack v defence was really beneficial because not only then couldyougettorecogniseit[patterns],butyou also get an understanding with your teammates…” (P2).
Central defenders expressed how game realistic practices are fundamental to developing pattern recognition skills. This is supported by previous literature that has emphasised the importance of representative practice environments with high levels of action fidelity and task functionality to promote the coupling of perception and action (Broadbent et al, 2015). Moreover, previous literature has stressed the importance of designing practice that involves repetition of monitoring ‘off the ball’ opposition offensive movements whilst subsequently engaging with ball location in connection with teammates positioning (Williams & Davids, 1998). Coach 4 shared:
“… we can go and stand behind our defenders to see what they see, and we can really, we can talk to them, we can slow it down, we can give them more detail about their movements” (C4).
Coaches emphasises the benefits of 'coaching through the game'. For example, standing behind central defenders in practice activities and talking to them can help develop their ability to recognise patterns of play emerging infront of them, with coaches being able to direct their attention to the most information-rich areas.
Central defenders and coaches also commented on the use of small-sided games ('playing form activities', see Ford et al., 2010) compared with 11v11 games, macro patterns, micro patterns, and pattern recognition principles:
“… it [pattern recognition] is challenging if you're training on a smaller pitch toreally hit homepattern recognition, because the game the distances in a game compared to playing on perhaps like a small six a side pitch or an eight a side pitch is completely different” (P1).
Player 1’s perspective is supported by research into decisionmaking in soccer which has identified how small-sided games fail to recreate the demands of a 11v11 match due to pitch size constraints resulting in players not performing frequent 11v11 actions such as long passes (O'Connor et al., 2017). On the contrary, previous research has found similarities between 11-vs11 and small-sided games with one particular investigation finding that 7-vs-7 small-sided games were faster paced, yet representative of 11-vs-11 games with regards to the performance indicators measured (Bergkamp et al., 2020). More specifically, the study found 7-vs-7 small-sided games were representative of 11-vs-11 games with regards to actions performed, excluding aerial duels (Bergkamp et al., 2020). Coach 1 commented:
“… you've got to solve the problem you've got in a small-sided game in a larger sided game. Ultimately, you've got to recognise patterns of play, and where the ball is going to travel to, etc.” (C1).
Coach 1 challenges suggestions that small-sided games in practice may hinder a player’sability to recognise full-sided game patterns (Runswick et al., 2021; Hope et al., 2024). Coach 3 highlights how "The eleven a side game is just a whole passage of 1v1, 2v2, 3v3, 4v4, up to your five asides … trying to get players sometimes to recognise, yes, this is 11 v 11, but in this moment, for this next 5-seconds, it's a 2 v 1, it's a 3 v 2". By implication, soccer players and coaches have a collective understanding that pitch distances and player numbers influence pattern recognition, yet micro-patterns (i.e., 2v2) emerge during small-sided games that directly translate into moments in a full-sided game (macro pattern). In light of these findings, small-sided games may have important implications for the transfer of skills from practice to competition (Broadbent et al., 2015).
4. Discussion
The present study aimed to qualitatively investigate the mechanisms and processes underpinning how players recognise patterns in soccer, and to investigate the importance of pattern recognition in competition and practice environments. Participants were asked to provide their own definition of ‘game reading’ (anticipation) in soccer. Their understanding appeared to encompass abilities of pattern recognition and visual exploratory activity (‘scanning’) to support their ability to anticipate successfully. Research has highlighted the interaction of different perceptual-cognitiveskillsin various situations (Roca& Williams, 2016). Therefore, anticipation, which has been captured through
the more commonly used term in soccer ‘game reading’, may include whatplayers are currently ‘reading’(i.e., what is currently unfolding on the pitch) and may also be ‘reading’ what is likely to happen next. To summarise the key findings, both players and coaches explained the importance of central defenders recognising ‘danger’ (i.e., runs in behind the defence from central attacking players, or a lack of pressure on an opposing midfield player on theball) which appeared linked to the distances between the defender and the ball. Based upon the experiences of both players and coaches, it is worth noting that effective ‘scanning’ behaviours can underpin this ability to recognise danger and distance to the ball.
Coaches and central defenders further emphasised the importance of experience and exposure to frequently occurring patterns of play which can develop anticipatory skills, positional awareness, and recognition of opposition team tendencies. As evidence, Player 5 shared “I don't think you'd do anything without having that pattern recognition, even making a tackle, you're recognising something, you're either recognising a striker’s movement a weight of pass to come win the ball”. These findings can be interpreted in relation to the theory of long-term working memory (Ericsson & Kintsch, 1995). Research has stated experts develop highly specialised knowledge structures through extended domain specific practice that enable them to identify structure and familiarity (North & Williams, 2019). Therefore, as shown by the experiential knowledge presented in this study, skilled central defenders may possess complex cognitive knowledge structures, which could have developed through repetition of extended, and specific practice. As a result, players may beabletorapidlyretrievetaskspecificinformationfromtheir long-term working memory (Ericsson & Kintsch, 1995) to anticipate future actions quickly and accurately, as well as perceive and recognise unfolding patterns of play.
Lastly, players and coaches frequently referred to how team dynamics influence pattern recognition and how the opposition team’s tactical set-up (e.g., style of play, formation and inpossession tendencies) can provide valuable information to support the pattern recognition process. Participants highlighted how a central defender’s role would involve frequently communicating to teammates to organise players into suitable positions on the pitch. These findings add to previous research which have highlighted the important role communication plays in team sports (Klatt & Smeeton, 2020; Riches et al., 2021), and how an opposition team’s in-possession tendencies may influence a player’s ability to recognise patterns of play. In light of these findings, there is a considerable lack of research that exists on understanding how team dynamics influence processes such as pattern recognition. We therefore encourage future research to investigate the role team dynamics (e.g., on-pitch verbal communication, opposition team tactical set-up and in-game coach instruction) play on soccer players’ ability to recognise patterns of play.
4.1. Recommendations for practice design
Based on current findings, numerous implications are proposed for designing practice. Firstly, coaches are recommended to design practice activities that require central defenders to continually evaluate their defensive position. During these
activities, coaches are encouraged to introduce game-related conditions that promote communication between central defenders and the rest of their defensive unit whilst ensuring players are continually mapping the location of the ball. Additionally, designing activities on a full-size pitch may facilitate the representativeness of practice which would enable central defenders to gain experience visually exploring their environment to locate ‘off the ball’ movements of opposition players and teammates. Techniques such as shadow coaching and questioning may aid the development of positional awareness, however coaches should avoid overloading players through constant instructions (i.e., 'over-coaching'), which has been reported to restrict guided discovery and may result in lower retention of skills, and performance breakdown under pressure (Ford et al., 2010; O'Connor et al., 2017). Secondly, small-sided games are encouraged to expose players to a high repetition of frequently occurring micro-patterns (i.e., 2v2, 3v3) that central defenders may encounter in dynamic instances in 11v11 matches. Thirdly, modified games (i.e., attack vs defence) in practice in which the opposition team replicates future oppositions’ style of play and attacking movement patterns could provide central defenders with a priori knowledge to develop game reading and pattern recognition skills.
4.2. Future research directions
Despite the novelty of the current study interviewing both male and female skilled central defenders and experienced male coaches, no female soccer coaches were sampled. Previous literature has referred to a lack of female soccer coaches working within elite soccer, with only 7% of all soccer coaches being women (De Haan, 2020). Therefore, as women’s soccer grows in popularity and the number of female coaches increase, future research should aim to prioritise the sampling of elite female soccer coaches. Furthermore, with the current study sampling bothmaleandfemalesoccerplayerstheremaybedifferentpattern recognition requirements across male and female soccer players. Previous research has found differences in game structure across male and female soccer matches (Tenga et al., 2015), which could impact upon how players recognise patterns of play. However, more recently, research by O’Brien-Smith et al. (2020) has shown no differences in decision making between male and female youth soccer players suggesting pattern recognition processes may not differ. From the themes that emerged from the current study, no clear differences were found between skilled male and female soccer players in how pattern recognition processes were used to support skilled performance. As authors we are conscious of the bias towards the sampling of male participants in scientific research and the lack of research investigating female participants (Williams et al., 2020) and so there is a critical need to undertake more research with female participants (Elliott-Sale et al., 2021). The current investigation relied upon the accuracy of participants’ recall ability to disclose previous experiences, which must be considered when evaluating the findings (see Hopwood, 2015). Semi-structured interviews were adopted which only focus on the conscious pick up of visual information processed from the environment (see Morris-Binelli, 2020), failing to identify the sub-conscious information used during competition. However, some participants alluded to the subconscious processing of
information as a result of experience. In light of the findings, future investigations should aspire to better understand the transferability of pattern recognition and game reading skillsfrom practice into competition. Enhancements made in virtual technology in soccer (see Wirth et al., 2018) and adopting more in-situ designs (see van Maarseveen et al., 2018) provide players with more representative viewing perspectives allowing for increased repetitions, variation in practice conditions, and rehearsal of scenarios that are challenging to reproduce in traditional practice (Gray, 2019). Whilst a representative viewing perspective allows for realistic information from opponents, there should also be consideration around interacting with team-mates (e.g., the defensive unit) when practising relevant scenarios (Janssen et al., 2023).
Previous research has suggested that practice in virtual reality environments may support one’s ability to ‘read the game’ at a more unconscious level (Ferrer et al., 2020). An interesting approach therefore is to utilise the emerging technology of 360° video to present soccer central defenders with real-world video footage of emerging patterns of play from a first-person perspective presented through a head-mounted display. This display offers innovative alternatives to traditional video footage, which is typically presented from a third person perspective, and allows the player to scan their environment (Lindsey et al., 2023). This approach provides unique opportunities to not only assess soccer players decision-making ability (see Honer et al., 2023), but to understand the relative importance of pattern recognition to game reading processes from a first-person perspective which could have direct implications for developing practice design (Musculus et al., 2021). Therefore, future research should combine both qualitative and quantitative findings to gain a more comprehensive understanding of pattern recognition processes which can then be developed to support the transfer of these skills into competition.
5. Conclusion
To conclude, this study is one of the first to provide qualitative data concerning pattern recognition and game reading in soccer, based upon the experiential knowledge of skilled central defenders and experienced coaches. Skilled central defenders continually utilise multiple perceptual-cognitive skills to position themselves in optimal locations on the pitch to enable the recognition of emerging patterns of play. Coaches provided insights into current practice design and the importance of representative practice to develop pattern recognition processes. Therefore, coaches should make greater efforts to incorporate more representative game-like situations during practice to allow players to become familiar with pitch distances and the identification of macro and micro patterns. To summarise, skilled central defenders recognise danger by player to ball distances and interactive movements of opposition central attacking players, which is linked to experience, sources of information, positional awareness, and organisation. Therefore, the study provides a novel insight into how skilled central defenders read the game and perceive familiarity in emerging patterns of play to produce high levels of performance in dynamic sporting environments.
Conflict of Interest
The authors declare no conflict of interests.
Acknowledgment
We would like to acknowledge the six skilled players and seven experienced coaches who participated in this project. We greatly appreciated both your time and expertise
References
Abernethy, B., Baker, J., & Côté, J. (2005). Transfer of pattern recall skills may contribute to the development of sport expertise. AppliedCognitivePsychology: TheOfficialJournal of the Society forApplied Research in Memory and Cognition, 19(6), 705–718. https://doi.org/10.1002/acp.1102
Abernethy,B.,&Russell,D.G.(1987).Expert-novicedifferences in an applied selective attention task. Journal of Sport and Exercise Psychology, 9(4), 326–345. https://doi.org/10.1123/jsp.9.4.326
Allard, F., Graham, S., & Paarsalu, M. E. (1980). Perception in sport: Basketball. Journal of Sport and Exercise Psychology, 2(1), 14–21. https://doi.org/10.1123/jsp.2.1.14
Bergkamp, T. L., den Hartigh, R. J., Frencken, W. G., Niessen, A. S. M., & Meijer, R. R. (2020). The validity of small-sided games in predicting 11-vs-11 soccer game performance. PLoS ONE, 15(9), 1–18 https://doi.org/10.1371/journal.pone.0239448
Bertram, R., Helle, L., Kaakinen, J. K., & Svedström, E. (2013). The effect of expertise on eye movement behaviour in medical image perception. PLoS ONE, 8, 1–15 http://dx.doi.org/10.1371/journal.pone.0066169
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101.
Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic analysis. Qualitative Research in Sport, Exercise and Health, 11(4), 589–597. https://doi.org/10.1080/2159676X.2019.1628806
Braun, V., Clarke, V., & Weate, P. (2016). Using thematic analysis in sport and exercise research. In B. Smith, & A. C. Sparkes (Eds.), Routledge Handbook of Qualitative Research in Sport and Exercise (pp. 191–205). Routledge.
Broadbent, D. P., Causer, J., Williams, A. M., & Ford, P. R. (2015). Perceptual-cognitive skill training and its transfer to expert performance in the field: Future research directions. European Journal of Sport Science, 15(4), 322–331. https://doi.org/10.1080/17461391.2014.957727
Burke, S. (2016). Rethinking ‘validity’ and ‘trustworthiness’ in qualitative inquiry: How might we judge the quality of qualitativeresearchinsportandexercisesciences?In B.Smith, & A.C. Sparkes (Eds.), Routledge Handbook of Qualitative Research in Sport and Exercise (pp. 330–341). Routledge. Carboch, J., Brenton, J., Reischlova, E., & Kocib, T. (2023). Anticipatory information sources of serve and returning of elite professional tennis players: A qualitative approach. International Journal of Sports Science & Coaching, 18(3), 761–771. https://doi.org/10.1177/17479541221092345
Feist et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 20-32 (2024)
Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4(1), 55–81. https://doi.org/10.1016/0010-0285(73)90004-2
De Haan, D. M. (2020). Levelling the playing field for women coaches in football. Infographic based on FIFA 2020 survey. The Hague University of Applied Sciences.
Den Hartigh, R. J., Van Der Steen, S., Hakvoort, B., Frencken, W. G., & Lemmink, K. A. (2018). Differences in game reading between selected and non-selected youth soccer players. Journal of Sports Sciences, 36(4), 422–428. https://doi.org/10.1080/02640414.2017.1313442
Eldridge, D., Pocock, C., Pulling, C., Kearney, P., & Dicks, M. (2023). Visual exploratory activity and practice design: Perceptions of experienced coaches in professional football academies. International Journal of Sports Science & Coaching, 18(2), 370–381 https://doi.org/10.1177/17479541221122412
Elliott-Sale, K. J., Minahan, C. L., de Jonge, X. A. J., Ackerman, K. E., Sipilä, S., Constantini, N. W., Lebrun, C. M., & Hackney, A. C. (2021). Methodological considerations for studies in sport and exercise science with women as participants: A working guide for standards of practice for research on women. Sports Medicine, 51(5), 843–861. https://doi.org/10.1007/s40279-021-01435-8
Ericsson, K. A., & Kintsch, W. (1995). Long-term working memory. Psychological Review, 102(2), 211–245. https://doi.org/10.1037/0033-295X.102.2.211
Ericsson, K. A., Nandagopal,K., &Roring,R. W. (2009).Toward a science of exceptional achievement: Attaining superior performance through deliberate practice. Annals of the New York Academy of Sciences, 1172(1), 199–217. https://doi.org/10.1196/annals.1393.001
Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100(3), 363–406.
Ferrer, R. C. D., Shishido, H., Kitahara, I., & Kameda, Y. (2020). Read-the-game: System for skill-based visual exploratory activity assessment with a full body virtual reality soccer simulation. PLoS ONE, 15(3), 1–29 https://doi.org/10.1371/journal.pone.0230042
Ford, P. R., Yates, I., & Williams, A. M. (2010). An analysis of practice activities and instructional behaviours used by youth soccer coaches during practice: Exploring the link between science and application. Journal of Sports Sciences, 28(5), 483–495. https://doi.org/10.1080/02640410903582750
Gibson, J. J. (1979). Theecologicalapproachtovisualperception. Houghton Mifflin.
Gonçalves, B., Coutinho, D., Exel, J., Travassos, B., Lago, C., & Sampaio, J. (2019). Extracting spatial-temporal features that describe a team match demands when considering the effects of the quality of opposition in elite football. PLoS ONE, 14(8), 1–20. https://doi.org/10.1371/journal.pone.0221368
Gray, R. (2019). Sports training technologies: Achieving and assessing transfer. In N.J. Hodges, & M.A. Williams (Eds.), Skill Acquisition in Sport (pp. 203–219). Routledge. Greenwood, D., Davids, K., & Renshaw, I. (2014). Experiential knowledge of expert coaches can help identify informational constraints on performance of dynamic interceptive actions. Journal of Sports Sciences, 32(4), 328–335. https://doi.org/10.1080/02640414.2013.824599
Hope, E. R., Patel, K., Feist, J., Runswick, O. R., & North, J. S. (2024). Examining the importance of local and global patterns for familiarity detection in soccer action sequences. Perception, 53(3),149–162. https://doi.org/10.1177/03010066231223825
Hodges, N. J., Wyder-Hodge, P. A., Hetherington, S., Baker, J., Besler, Z., & Spering, M. (2021). Topical review: Perceptualcognitive skills, methods, and skill-based comparisons in interceptive sports. Optometry and Vision Science, 98(7), 681–695. https://doi.org/10.1097/OPX.0000000000001727
Hopwood, M. J. (2015). Issues in the collection of athlete training histories. In J. Baker, & D. Farrow (Eds.), Routledge Handbook of Sport Expertise (pp. 156–165). Routledge. Höner, O., Dugandzic, D., Hauser, T., Stügelmaier, M., Willig, N., & Schultz, F. (2023). Do you have a good all-around view? Evaluation of a decision-making skills diagnostic tool using 360° videos and head-mounted displays in elite youth soccer. Frontiers in Sports and Active Living, 5, 1–11. https://doi.org/10.3389/fspor.2023.1171262
Janssen, T., Müller, D., & Mann, D. L. (2023). From natural towards representative decision making in sports: A framework for decision making in virtual and augmented environments. Sports Medicine, 53(10), 1851–1864.
Jewell, S., Murphy, C. P., Pocock, C., & North, J. S. (2022). Exploring how the experiences of English youth football coaches have shaped their approach to player learning. The Journal of Sport and Exercise Science, 6(3), 179–190. https://doi.org/10.36905/jses.2022.03.04
Jordet, G., Aksum, K. M., Pedersen, D. N., Walvekar, A., Trivedi, A.,McCall,A.,Ivarsson,A.,&Priestley,D.(2020).Scanning, contextual factors, and association with performance in English Premier League soccerers: An investigation across a season. Frontiers in Psychology, 11, 1–16. https://doi.org/10.3389/fpsyg.2020.553813.
Klatt, S., & Smeeton, N. J. (2020). Visual and auditory information during decision making in sport. Journal of Sport and Exercise Psychology, 42(1), 15–25. https://doi.org/10.1123/jsep.2019-0107
Levi, H. R., & Jackson, R. C. (2018). Contextual factors influencing decision making: Perceptions of professional soccer players. Psychology of Sport and Exercise, 37, 19–25. https://doi.org/10.1016/j.psychsport.2018.04.001
Lindsay, R., Spittle, S., & Spittle, M. (2023). Skill adaption in sport and movement: Practice design considerations for 360° VR. Frontiers in Psychology, 14, 1–5. https://doi.org/10.3389/fpsyg.2023.1124530
Malterud, K., Siersma, V. D., & Guassora, A. D. (2016). Sample size in qualitative interview studies: Guided by information power. Qualitative Health Research, 26, 1753–1760. https://doi.org/10.1177/1049732315617444
Marteniuk, R. G. (1976). Information Processing in Motor Skills. Holt, Rinehart, and Winston.
McLean, S.,Salmon,P.M., Gorman,A.D.,Dodd,K.,&Solomon, C. (2021). The communication and passing contributions of playing positions in a professional soccer team. Journal of Human Kinetics, 77, 223–234. https://doi.org/10.2478/hukin2020-0052
Morris-Binelli, K., van Rens, F. E., Müller, S., & Rosalie, S. M. (2020). Psycho-perceptual-motor skills are deemed critical to save the penalty corner in international field hockey.
Feist et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 20-32 (2024)
Psychology of Sport and Exercise, 51, 1–9. https://doi.org/10.1016/j.psychsport.2020.101753
Murphy, C. P., Jackson, R. C., & Williams, A. M. (2018). The role of contextual information during skilled anticipation. QuarterlyJournalofExperimentalPsychology,71(10),2070–2087. https://doi.org/10.1177/1747021817739201
Musculus, L., Bäder, J., Sander, L., & Vogt, T. (2021). The influence of environmental constraints in 360° videos on decision making in soccer. Journal of Sport and Exercise Psychology, 43(5), 365–374. https://doi.org/10.1123/jsep.2020-0166
North, J. S., Hope, E., & Williams, A. M. (2016). The relative importance of different perceptual-cognitive skills during anticipation. Human Movement Science, 49, 170–177. https://doi.org/10.1016/j.humov.2016.06.013
North, J. S., Hope, E., & Williams, A. M. (2017). Identifying the micro-relations underpinning familiarity detection in dynamic displays containing multiple objects. Frontiers in psychology, 8, 1–8. https://doi.org/10.3389/fpsyg.2017.00963
North, J. S., Ward, P., Ericsson, A., & Williams, A. M. (2011). Mechanismsunderlying skilledanticipationandrecognitionin a dynamic and temporallyconstrained domain. Memory, 19(2), 155–168. https://doi.org/10.1080/09658211.2010.541466
North, J. S., & Williams, A. M. (2019). Familiarity detection and pattern perception. In M. Williams & R. Jackson (Eds.). Anticipation and Decision Making in Sport (pp. 25–42). Routledge.
North, J. S., Williams, A. M., Hodges, N., Ward, P., & Ericsson, K. A. (2009). Perceiving patterns in dynamic action sequences: Investigating the processes underpinning stimulus recognition and anticipation skill. Applied Cognitive Psychology: The Official Journal of the Society for Applied Research in Memory and Cognition, 23(6), 878–894. https://doi.org/10.1002/acp.1581
Nowell, L. S., Norris, J. M., White, D. E., & Moules, N. J. (2017). Thematic analysis: Striving to meet the trustworthiness criteria. International Journal of Qualitative Methods, 16(1), 1–13. https://doi.org/10.1177/1609406917733847
O’Brien-Smith, J., Bennett, K. J., Fransen, J., & Smith, M. R. (2020). Same or different? A comparison of anthropometry, physical fitness and perceptual motor characteristics in male and female youth soccer players. Science and Medicine in Football, 4(1), 37–44. https://doi.org/10.1080/24733938.2019.1650197
O’Connor, D., Larkin, P., & Williams, A. M. (2017). What learning environments help improve decision-making? Physical Education and Sport Pedagogy, 22(6), 647–660. https://doi.org/10.1080/17408989.2017.1294678
Pocock, C., Bezodis, N. E., Davids, K., Wadey, R., & North, J. S. (2020). Understanding key constraints and practice design in Rugby Union place kicking: Experiential knowledge of professional kickers and experienced coaches. International Journal of Sports Science & Coaching, 15(5–6), 631–641. https://doi.org/10.1177/1747954120943073
Riches, K. M., Murphy, C. P., & Broadbent, D. P. (2021). Skillbased differences in decision time when responding to verbal information from intrateam communication in lacrosse. Journal of Expertise, 4(4), 365–374
Roca, A., Ford,P. R., McRobert, A. P., &Williams, A. M. (2011). Identifying the processes underpinning anticipation and
decision-making in a dynamic time-constrained task. Cognitive Processing, 12(3), 301–310. https://doi.org/10.1007/s10339-011-0392-1
Roca, A., Ford,P. R., McRobert, A. P., &Williams, A. M. (2013). Perceptual-cognitive skills and their interaction as a function of task constraints in soccer. Journal of Sport and Exercise Psychology, 35(2), 144–155. https://doi.org/10.1123/jsep.35.2.144
Roca, A., & Williams, A. M. (2016). Expertise and the interaction between different perceptual-cognitive skills: Implications for testing and training. Frontiers in Psychology, 7, 1–4. https://doi.org/10.3389/fpsyg.2016.00792
Runswick, O., Hope, E., Feist, J., Patel, K., & North, J. S. (2021). Recognition of micro-relations in the context of full game patterns in soccer. Journal of Sport and Exercise Psychology, 43, S45–S45
Russo, M. B., Kendall, A. P., Johnson, D. E., Sing, H. C., Thorne, D.R.,Escolas, S.M., Santiago,S.,Holland,D.A.,Hall, S.W., & Redmond, D. P. (2005). Visual perception, psychomotor performance, and complex motor performance during an overnight air refuelling simulated flight. Aviation, Space, and Environmental Medicine, 76(7), 92–103.
Seifert, L., Button, C., & Davids, K. (2013) Key properties of expert movement systems in sport: An ecological dynamics perspective. Sports Medicine, 43(3), 167–178.
Smith, B., & Caddick, N. (2012). Qualitative methods in sport: A concise overview for guiding social scientific sport research. Asia Pacific Journal of Sport and Social Science, 1(1), 60–73. https://doi.org/10.1080/21640599.2012.701373
Smith, M., & Cushion, C. J. (2006). An investigation of the ingame behaviours of professional, top-level youth soccer coaches. Journal of Sports Sciences, 24(4), 355–366. https://doi.org/10.1080/02640410500131944
Smith, B., & McGannon, K. R. (2018). Developing rigor in qualitative research: Problems and opportunities within sport and exercise psychology. International Review of Sport and Exercise Psychology, 11(1), 101–121. https://doi.org/10.1080/1750984X.2017.1317357
Smith, B., & Sparkes, A. C. (2016). Interviews: Qualitative interviewing in the sport and exercise sciences. In B. Smith, & A. C. Sparkes (Eds.), Routledge Handbook of Qualitative Research in Sport and Exercise (pp. 103-124) Routledge. Sparkes, A. C., & Smith, B. (2014). Qualitative research methods in sport, exercise and health: From process to product Routledge.
Strafford, B. W., Davids, K., North, J. S., & Stone, J. A. (2021). Designing Parkour-style training environments for athlete development: Insights from experienced Parkour Traceurs. Qualitative Research in Sport, Exercise and Health, 13(3), 390–406. https://doi.org/10.1080/2159676X.2020.1720275
Tracy, S. J. (2010). Qualitative quality: Eight “big-tent” criteria for excellent qualitative research. Qualitative Inquiry, 16(10), 837–851. https://doi.org/10.1177/1077800410383121
Tenga, A., Zubillaga, A., Caro, O., & Fradua, L. (2015). Explorative study on patterns of game structure in male and female matches from elite Spanish soccer. International Journal of Performance Analysis in Sport, 15(1), 411–423. https://doi.org/10.1080/24748668.2015.11868802
van Maarseveen, M. J., Oudejans, R. R., Mann, D. L., & Savelsbergh, G. J. (2018). Perceptual-cognitive skill and the
Feist et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 20-32 (2024)
in-situ performance of soccer players. Quarterly Journal of Experimental Psychology, 7(2), 455–470. https://doi.org/10.1080/17470218.2016.1255236
Ward,P.,&Williams,A.M.(2003).Perceptualandcognitiveskill development in soccer: The multidimensional nature of expert performance. Journal of Sport & Exercise Psychology, 25(1), 93–111. https://doi.org/10.1123/jsep.25.1.93
Williams, A. M., & Davids, K. (1998). Visual search strategy, selective attention, and expertise in soccer. Research Quarterly for Exercise and Sport, 69(2), 111–128. https://doi.org/10.1080/02701367.1998.10607677
Williams, A. M., Ford, P. R., & Drust, B. (2020). Talent identification and development in soccer since the millennium. Journal of Sports Sciences, 38(11–12), 1199–1210. https://doi.org/10.1080/02640414.2020.1766647
Williams, A. M., Ford, P. R., Eccles, D. W., & Ward, P. (2011). Perceptual‐cognitive expertise in sport and its acquisition: Implications for applied cognitive psychology. Applied Cognitive Psychology, 25(3), 432–442. https://doi.org/10.1002/acp.1710
Williams, A. M., Heron, K., Ward, P., & Smeeton, N. J. (2005). Using situational probabilities to train perceptual and cognitive skill in novice soccer players. In T. Reilly, J. Cabri,
& D. Araùjo (Eds.), Science and football V, (pp. 337–340) Routledge
Williams, A. M., Hodges, N. J., North, J. S., & Barton, G. (2006). Perceiving patterns of play in dynamic sport tasks: Investigating the essential information underlying skilled performance. Perception, 35(3), 317–332. https://doi.org/10.1068/p5310
Williams, A. M., & Jackson, R. C. (2019). Anticipation in sport: Fifty years on, what have we learned and what research still needs to beundertaken? PsychologyofSportandExercise,42, 16–24. https://doi.org/10.1016/j.psychsport.2018.11.014
Williams, A. M., North, J. S., & Hope, E. R. (2012). Identifying the mechanisms underpinning recognition of structured sequences of action. The Quarterly Journal of Experimental Psychology, 65(10), 1975–1992. https://doi.org/10.1080/17470218.2012.678870
Wirth, M., Gradl, S., Poimann, D., Schaefke, H., Matlok, J., Koerger, H., & Eskofier, B. M. (2018). Assessment of perceptual-cognitive abilities among athletes in virtual environments: Exploring interaction concepts for soccer players. Proceedings of the 2018 Designing Interactive Systems Conference, 1013–1023. https://doi.org/10.1145/3196709.3196780
The Journal of Sport and Exercise Science, Vol. 8, Issue 1, 33-42 (2024)
www.jses.net
Effects of three weeks of aerobic blood flow restriction completed before daily training on running performance and resting hemodynamic measures in trained runners
Alexander H. K. Montoye1*, Jackson T. Nordbeck1, Benjamin A. Cox2, Munni Begum3, Drew M. Lazar3 , Jennifer R. Vranish1
1Department of Integrative Physiology and Health Science, Alma College, USA
2Cox Sports Medicine and Orthopedic Surgery, USA
3Department of Mathematical Sciences, Ball State University, USA
Received: 16.12.2023
Accepted: 17.07.2024
Online: 24.09.2024
Keywords:
Limb occlusion pressure
Arterial occlusion pressure
Ischemic preconditioning
Ergogenic aid
Sport science
Our study investigated the effects of a daily, pre-training aerobic blood flow restriction (BFR) protocol on 2.4 km running time trial performance and resting cardiovascular measures and hemodynamics. Runners on a collegiate cross country team (n = 13, 61.5% female, aged 18 – 23 years) participated in a two-group randomized crossover design, completing a maximal 2.4 km running time trial and resting cardiovascular (heart rate, brachial blood pressure) and hemodynamic (popliteal artery: diameter, velocity, blood flow, shear rate) measures at baseline. One group (Group B) performed BFR for 10 – 15 minutes at 65% of limb occlusion pressure while engaging in light-intensity walking prior tocompletingateam-prescribedrunningworkoutonatleastfivedays/weekforthreeweeks, while another group (Group A) completed the same light-intensity walking and same running workout but without BFR. After three weeks, the running and resting cardiovascular and hemodynamic measures were repeated. Following a 2- to 3-day washoutperiod,the groupsreversed,with GroupAintheinterventionconditionandGroup B in the control condition for three weeks. Final running and resting cardiovascular and hemodynamic measures were then taken. There were no significant treatment effects for any measured outcome variables. However, changes in 2.4 km running time in the interventionconditionexceededthesmallestworthwhilechange(-20.8±24.7seconds,95% CI [-34.2, -7.3]), with no improvement in the control condition (-7.2 ± 18.7 seconds, 95% CI [-17.4, 3.0]). Daily, pre-training aerobic BFR at 65% of limb occlusion pressure may elicit a meaningful, favorable change in 2.4 km time trial performance in trained collegiate runners. If future studies confirm our findings, our protocol may be appealing for use in field-based settings.
1. Introduction
Blood flow restriction (BFR) is a technique that uses restrictive cuffs, wraps, or similar equipment on the arm or leg to reduce blood flow during exercise. Initially popularized in resistance training, BFR has gained significant attention for its potential to enhance muscle hypertrophy and strength gains with lower exercise intensities compared to traditional high-load training (Hughes et al., 2017; Pignanelli et al., 2021; Scott et al., 2015;
Slysz et al., 2016). BFR techniques have also been incorporated into low-intensity aerobic exercises such as cycling, walking, and rowing. Evidence in this research area suggests that, in nonathletes, aerobic BFR training may induce a range of physiological adaptations including improved aerobic capacity (VO2max), muscular strength and endurance, and vascular function (Abe et al., 2006; 2010; de Oliveira et al., 2016). One systematic review found evidence of superior VO2max and aerobic performance outcomes with aerobic BFR training
*Corresponding Author: Alexander H. K. Montoye, Alma College, USA, montoyeah@alma.edu
(Bennett & Slattery, 2019), and another review found greater muscular fitness (strength, endurance) benefits of aerobic BFR training (de Lemos Muller et al., 2024). Of note, both of these reviews focused on healthy non-athletes and comparing BFR training to absolute intensity-matched training without BFR. However, endurance athletes differ from non-athletes in many ways and, therefore, may not experience the same adaptations with BFR training.
Peak endurance exercise performance is determined by a number of factors such as VO2max, muscle function, exercise economy, and lactate threshold. Vascular function is an important determinant of oxygen and nutrient delivery to tissues as well as lactate clearance and, therefore, influences many of these aforementioned factors. Two studies in elite rowers found improvements in VO2max when incorporating BFR during lowintensity rowing training compared to control groups completing similar training without BFR (Held et al., 2020; 2024). However, one of these studies also assessed 1-repetition maximum squat performance, finding no difference between intervention and control groups (Held et al., 2020), and the other study also examined peak power output and time trial performance and found no difference in these variables between intervention and control groups (Held et al., 2024) A similar study in elite rowers found improvements in VO2max and time trial performance with four weeks of aerobic BFR training but lacked a control group in order to determine if and how much the BFR element played in the performance improvements (Thompson et al., 2024) A study in masters level cyclists foundthat supplementing normal training with aerobic BFR enhanced cycling time trial performance, but this could have been due to the increased training volume in the BFR group compared to the control group (Tangchaisuriya et al., 2022). In contrast, a study which utilized sprint interval training in trained cyclists found similar improvement in time trial performance regardless of whether the training was conducted with or without BFR (Giovanna et al., 2022) Finally, two related studies in endurance runners found improvements in VO2max, running performance, and muscular strength and endurance with an eight-week running BFR training program when compared to an absolute intensity-matched control group without BFR (Chen, Hsieh, Ho, Ho, et al., 2022; Chen, Hsieh, Ho, Lin, et al., 2022) In many of these studies, the BFR training was done in place of normal training rather than as a supplement to a typical training routine. Additionally, the studies which specified when they were conducted took place in athletes’ off-season, offering little insight as to whether BFR training can be beneficial when supplemented into in-season training. In summary, the available evidence provides intriguing but limited and mixed evidence as to if and how training with BFR may affect exercise performance in athletes.
When evaluating the current evidence determining the effectiveness of aerobic BFR training in enhancing endurance athletes’ exercise performance, a number of noteworthy limitations exist. First, much of the current work has evaluated aerobic BFR training to absolute intensity-matched exercise without BFR, with these protocols serving as the main training stimulus. Studies examining the effects of resistance training with BFR demonstrate superior training adaptations with BFR when compared to absolute intensity-matched exercise without BFR (Pignanelli et al., 2021; Scott et al., 2015; Wortman et al., 2021),
whereas resistance training with BFR typically produces smaller training adaptations than high-load training without BFR (Lixandrao et al., 2018). In both of these cases, magnitude and directionality of training adaptations with aerobic BFR likely mirror those of resistance BFR training. Therefore, from an athlete training perspective it is of more value to evaluate the use of aerobic BFR training as a supplement, rather than a replacement, for their normal training when seeking to optimize training adaptations. Second, the efficacy of BFR training during an athlete’s season (in contrast to use in the off-season) is unknown.
Finally, mixed evidence is available evaluating the effects of aerobic BFR on hemodynamic and vascular changes in endurance athletes. A review which evaluated evidence of resistance BFR training found superior changes in muscle capillarization compared to absolute intensity-matched training without BFR (Ferguson et al., 2021). Another review found some evidence of superior improvement in endothelial function but no difference in changes in resting heart rate or blood pressure with resistance BFR training compared to absolute intensity-matched training without BFR (Maga et al., 2023) However, both of these reviews included studies with primarily non-athlete populations. Given the potential for maladaptive changes to vascular function with BFR training noted in several recent reviews (da Cunha Nascimento et al., 2020; Spranger et al., 2015) and the lack of research examining such potential changes in athletes, more work is needed to determine potential changes in cardiovascular and hemodynamic function in athletic populations.
With these research gaps in mind, our study’s primarypurpose was to assess the effects of three weeks of almost daily walking with BFR prior to normal training on 2.4 km running time trial performance in collegiate cross country runners. Our secondary purpose was to assess the effects of this walking BFR program on resting cardiovascular and hemodynamic measures.
2. Methods
2.1
Participants
Males and females aged 18 – 23 years who were active members of a collegiate Division III cross country team and who reported 3 – 11 years of competitive endurance running experience were recruited for this study. Potential participants were excluded if they had a known cardiovascular, renal, or metabolic disease; were known tobe at increased risk of blood clotting; had anybone or muscle injuries in the lower limbs that prevented full participation in their training schedule; or had COVID-19 in the previous 4 weeks. Although 19 participants (9 male, 10 female) started the study, only 13 (5 male, 8 female) completed training sessions and all testing to be included in the final analysis. Of note, data collection was performed during the 7-week fall cross country season, and dropout from this study was due to injuries, illness, or other factors unrelated to study participation. No injuries or negative side effects due to the BFR protocol were noted. All participants in this study completed an informed consent form and were given the ability to ask questions prior to and during the study. ThisstudywasapprovedbytheAlmaCollege Institutional Review Board (IRB#: R_2VrKwENd78z7MHQ) prior to beginning testing.
2.2. Procedure
Participants completed several resting measures within a laboratory setting at three time points: baseline, mid-point, and post. For these assessments, participants were asked to abstain from food, caloric drinks, exercise, caffeine, and tobacco for at least three hours prior to laboratory visits. Upon arrival, participant height and weight were taken using a stadiometer (Seca GmbH & Co. KG., Hamburg, Germany) and scale (Tanita Corp., Tokyo, Japan), respectively, and age was self-reported. Thigh circumference was assessed at one third of the distance from the inguinal crease to the top of the patella for each thigh. Following 3 – 5 minutes of supine rest, resting heart rate and blood pressure were assessed using a Welch Allyn ProBP 3400 system (Hillrom, Skaneateles Falls, New York, USA), with the cuff placed on the left arm.
Next, resting blood flow through the popliteal artery of each leg while supine was determined by a researcher (JRV) with extensive training in ultrasound-based vascular measures (Vranish et al., 2017). For blood flow determination, a GE LOGICe ultrasound machine (GE Healthcare, Chicago, IL, USA) with a 9 MHz linear probe was used to measure the diameter and blood velocity in the popliteal artery. Using screen capture software (QuikTime Pro, Apple Inc., Cupertino, CA, USA), 30-second videos of blood flow in the popliteal artery were recorded, and Cardiovascular Suite (Quipu, Pisa, Italy) was used to obtain indices of average artery diameter and blood velocity during the measurement period, from whichtotalbloodflowwascalculatedusingthefollowingequations (Gnasso et al., 2001; Secomb, 2016):
direction), paved trail which was flat and had extensive tree cover to shield participants from the wind. Spotters ensured that participants reached the 1.2 km mark before turning around. The total time to complete the 2.4 km course, measured to the nearest second, was recorded for each participant by a trained research assistant. Participants were then allowed a self-paced cool-down.
Our study’s purpose was to assess the effects of pre-training, aerobic BFR on the aforementioned time trial and laboratory outcomes. To accomplish this purpose, our study used a 2 × 2 crossover design where, following baseline testing, participants were split into two groups. One group (Group A) received the control condition for the first three weeks (baseline to mid-point) followed by the intervention for the next three weeks (mid-point to post). The other group (Group B) received the intervention for the first three weeks (baseline to mid-point) followed by the control condition for the next three weeks (mid-point to post). Both groups had a 2- to 3-day washout period (depending on availability for performing laboratory testing following the time trial) between conditions. Groups were assigned at random, and this study design was chosen to account for seasonal, training, or learning effects. Figure 1 provides a schematic of the overall structure of the study.
Additionally, shear rate was calculated as:
To determine participants’ LOP, a lower body Personalized Tourniquet System (Delfi Medical Innovations, Inc., Vancouver, BC, Canada) was connected to a 11.5 cm width cuff, which was placed as proximally as possible on the thigh (choice of right vs. left as the first measure was randomized across participants), and the “Personalized Tourniquet Pressure” protocol was run on the Delfi system. Following 5 – 7 minutes of supine rest, the same procedure was conducted for the other thigh. The LOP measures obtained from the two legs were then averaged, and 65% of LOP was calculated for use in the aerobic BFR protocol. Past work has shown the efficacy of using cuff pressures of 50 – 75% of LOP (Montoye et al., 2023), and we selected a percentage near the middle of this range since participants would be in charge of inflating cuffs during the protocol, thus allowing for some user error while still remaining in the 50 – 75% of LOP range.
In addition to the resting laboratory measures, participants completed an outdoor, 2.4 km running time trial at baseline, midpoint, andpost. Priorto thetime trial, participants completeda 10minute dynamic stretching protocol and a self-paced 2.4 km joggingwarm-up.Then,participantscompleteda2.4kmtimetrial as quickly as possible on an out-and-back (1.2 km in each
In brief, all participants underwent baseline laboratory and time trial testing as described above. Then, the intervention group (Group B) was instructed to wear blood pressure cuffs (21 cm width; EverDixie, Dixie EMA Supply Co., Brooklyn, NY, USA) as proximally as possible on both thighs and to inflate them simultaneously to 65% of LOP. They were to walk at a selfdetermined light intensity (described as an intensity which was comfortable and did not cause them fatigue) for 10 – 15 minutes (while wearing the inflated cuffs) immediately prior to each running practice, for at least five days/week, over an approximately 3-week span. Participants needed to complete at least 10 minutes with the cuffs applied; because multiple participants were doing the cuff restriction simultaneously it was not always possible to control exact start/stop times, and some participants likely had a few extra minutes of restriction. Upon arrival at practice, participants removed the cuffs and completed a standardized, 10-minute dynamic warm-up that consisted of exercises such as high knees, skipping, and lunges performed at a controlled speed for 15 meters in a straight line. Next, participants completed theirnormalteam runningworkouts.Thecontrolgroup (Group A) completed the same 10 – 15 minutes light-intensity walk, standardized 10-minute dynamic warm-up, and team running workout but did not complete any BFR prior to practice. All walking and warm-up exercises were completed by Groups A and B at the same time. Following the three weeks, a mid-point set of time trial and laboratory tests were completed. Then, participants switched conditions, so Group A became the intervention group and Group B became the control group for the next three weeks. Following three weeks in the new conditions, a third and final set of laboratory and time trial tests (post) were performed. Research staff were present at all workouts to ensure compliance to the BFR protocol and answer questions from participants. Research staff also informally asked participants about potential pain/discomfort or fatigue from the BFR protocol that affected their workouts. Finally, participant race times throughout their racing season were collected from online Michigan Intercollegiate Athletic Association cross country race results.
2.3. Statistical approach
To investigate the effects of a low-intensity, aerobic BFR protocol performed at leastfive days/weekfor three weekson2.4km running time trial performance and resting cardiovascular and hemodynamic measures,weappliedlinearmixedeffectsmodels(LMEM;Singer& Willett, 2003) with a random intercept and unstructured covariance. Separately for each outcome of interest (2.4 km run time, resting heart rate, resting systolic and diastolic blood pressure, popliteal artery diameter, popliteal artery blood velocity, popliteal artery blood flow, popliteal artery shear rate), the LMEM includes period, treatment, and baseline measure of each outcome of interest.
In a 2 × 2 crossover study, the period effect implies that the effect of the same treatment received at two different periods may be different for a specific period (Lim & In, 2021). The LMEM for assessing the direct effect of the intervention includes
indicator variables for the intervention (treatment) and period effects after adjusting for the baseline outcome measure.
In addition tothe LMEM,effect size analyses were conducted, and effect sizes were considered negligible if < 0.20, small if 0.20 – 0.49, medium if 0.50 – 0.79, and large if ≥ 0.80 (Cohen, 1988). With an alpha level of p < 0.05, a desired power of 80%, and a sample size of 13, our study was powered to detect only large effect sizes (≥ 0.84). Thus, a smallest worthwhile change analysis was also conducted in order to detect smaller but potentially meaningful changes between conditions. For the smallest worthwhile change analysis, a meaningful difference was denoted as when the mean difference between trials exceeded 0.6 × standard deviation of the difference between trials (Buchheit, 2016; Marocolo et al., 2019). Analyses were conducted in R (R Core Development Team, Vienna, Austria) and Microsoft Excel 365 (Microsoft Corp., Redmond, WA, USA).
Figure 1: Flow chart of study procedure.
3. Results
Independent-samples t-tests confirmed that participants in Group A and B were not significantly different at baseline for any of the measured variables except for resting heart rate, which was significantly lower in Group B (Table 1).
Table 1: Baseline anthropometric, hemodynamic, and running time trial data for Group A and B.
Table 2: Results of 2 × 2 (group × time) linear mixed models analyses for run time, resting heart rate and blood pressure, and resting popliteal artery outcome variables.
Predictors Estimates 95% CI p
Run time
Intercept 90.59 [30.91, 150.28] 0.005* Seq[IC] -9.79 [-27.36, 7.78]
*
Run b 0.84 [0.75, 0.93] < 0.001*
Marginal ��2/Conditional ��2 = 0.947/0.962
Resting heart rate
Intercept 28.28 [1.35, 55.21] 0.041* Seq[IC] -8.10 [-14.65, -1.55] 0.018*
[-5.96, 3.86] 0.660
[-5.53, 4.29] 0.795
b 0.63 [0.21, 1.05] 0.005* Marginal ��2/Conditional ��2 = 0.576 / 0.656
Systolic blood pressure
37.77 [-31.71, 107.24] 0.269
[IC] -3.34 [-10.79, 4.11] 0.360 P 3.45 [-1.27, 8.17] 0.142
[-6.84 2.60] 0.359
b 0.67 [0.08, 1.25] 0.027*
Marginal ��2/Conditional ��2 = 0.363 / 0.610
Diastolic blood pressure
[23.82, 92.55] 0.002*
[IC] -4.92 [-9.10, -0.74] 0.023*
[-3.68, 3.68] 0.850
[-3.39, 3.35] 0.988
[-0.38, 0.65] 0.602
Marginal ��2/Conditional ��2 = 0.224/ 0.366
Popliteal artery blood flow
Notes:M =mean,SD=standard deviation, *Significantdifference from Group A (p < 0.05).
Results of the LMEM analyses are shown in Table 2, with data plotted by group and time shown in Figure 2 (also shown in table format in Table 3). For 2.4 km time trial completion time, there was a significant main effect with times decreasing from baseline. However, sequence, time, and treatment were not statistically significant. Resting heart rate and diastolic blood pressure were significantly lower in Group B than Group A, and there was a main effect for heart rate decreasing from baseline, but time and treatment were not statistically significant. For systolic blood pressure, popliteal artery blood flow velocity and shear rate, there were significant main effects but no significant effects of sequence, time, or treatment. For popliteal artery blood flow and popliteal artery diameter, there were no significant main effects nor effects of sequence, time, or treatment.
96.05 [54.35, 196.16] 0.002*
[IC] -24.29 [-73.77, 25.29] 0.317
-29.21 [-77.94, 19.52] 0.225
-15.73 [-64.47, 33.00] 0.507
0.42 [0.00, 0.84] 0.052
Marginal ��2/Conditional ��2 = 0.213/ 0.218
Popliteal artery diameter
Intercept 3.22 [1.15, 5.28] 0.004* Seq [IC] 0.27 [-0.51, 1.06] 0.478 P -0.12 [-0.40, 0.16] 0.394 T 0.14 [-0.14, 0.42] 0.308
DBP b 0.38 [-0.02, 0.77] 0.059
Marginal ��2/Conditional ��2 = 0.360/ 0.813
Popliteal artery blood flow velocity
Intercept 6.69 [-0.37, 13.75] 0.062 Seq [IC] -1.79 [-5.63, 2.04] 0.340
-2.20 [-5.79, 1.39] 0.216
-2.14 [-5.74, 1.45] 0.227 BV b 0.83 [0.19, 1.46] 0.014*
Marginal ��2/Conditional ��2 = 0.362/0.369
Popliteal artery shear rate
Intercept 86.27 [-17.84, 190.4] 0.099
Seq [IC] -17.19 [-84.77, 50.39] 0.601 P -35.39 [-95.83, 25.06] 0.235 T -40.96 [-101.4, 19.49] 0.172
SR b 0.97 [0.43, 1.50] 0.001*
Marginal ��2/Conditional ��2 = 0.488/NA
Notes: Seq, sequence; P, period (time); T, treatment; * p < 0.05.
Figure 2: Outcome variables at each time point, shown by group and condition. Error bars represent standard error. Dotted line indicates time where participants were in control condition. Solid line indicates time where participants were in intervention condition.
Table 3: Table with data from which Figure 2 was created.
Notes: M = mean, SD = standard deviation, ^Portion of the study in which each group received the intervention (Group B for the three weeks between Baseline and Mid-point testing, Group A for the three weeks between Mid-point and Post testing).
Results of the smallest worthwhile change analysis (Table 4) revealed no meaningful changes in resting heart rate, systolic blood pressure, or any of the popliteal artery outcomes despite medium effect sizes for all variables except popliteal artery total blood flow (small) and popliteal artery diameter (negligible). However, the change in diastolic blood pressure in the intervention condition exceeded the smallest worthwhile change (effect size = 0.62), with a non-significant trend toward greater change in the intervention group (drop of 3.6 ± 5.8 mmHg) compared to the control group (increase of 1.1 ± 4.1 mmHg).
Additionally, the change in time trial performance during the intervention condition exceeded the smallest worthwhile change (effect size = 0.62), with a non-significant trend toward lower running times in the intervention condition (improvement of 20.8 ± 24.7 seconds, which translates to 8.7 ± 10.3 seconds/km) compared to the control condition (improvement of 7.2 ± 18.7 seconds, or 3.0 ± 7.8 seconds/km). To give context to the extent of participant improvement in the time trial, participants’ fastest race of the season was 10.2 ± 7.8 seconds/km faster than their first race of the season.
Note: SWC = smallest worthwhile change, MD = mean difference, ^Indicates that difference between trials exceeded the smallest worthwhile change threshold.
4. Discussion
Our study evaluated the effects of an aerobic BFR protocol utilizing approximately 10 – 15 minutes restriction duration at 65% of LOP prior to daily running practices for three weeks in collegiate cross country runners during their fall season. Our results found no significant differences in any of the
cardiovascular, hemodynamic, or time trial variables between control and intervention conditions. However, the time trial results, while non-significant, did identify a potentially meaningful change in the favorable direction with BFR administration, with the time trial change exceeding the smallest worthwhile change and a medium effect size for differences from the control condition. When participants were in the BFR
Table 4: Smallest worthwhile change and effect size analyses for differences in outcome variables across conditions.
condition, point estimates of time trial performance improved by an average of approximately 8.7 seconds/km. For context, by examining published race reports, it was found that our study participants improved race times by 10.2 seconds/km throughout their seven-week season. If such an improvement could be maintained for full race distances (6.0 km for women, 8.0 km for men), our study suggests that the majority (~85%) of the improvement in race times across the season could be accounted for by changes in 2.4 km run times during the three weeks participants were in the intervention group.
Our findings addtoa mixed bodyof evidencewhen evaluating effects of aerobic BFR training on endurance performance in athletes. One meta-analysis evaluating aerobic BFR in athletes found effect sizes favoring a small additional gain in fitness with aerobic BFR training as compared to similar high-intensity training without BFR, but heterogeneity in findings and a small sample size kept this result from achieving statistical significance (Castilla-López et al., 2022) A recent review study indicated that aerobic BFR training can enhance VO2max, increase the lactate threshold, and possibly improve movement economy, although exact adaptations are likely to be protocol- and populationspecific (Smith et al., 2022). Many of the previously-conducted studies match absolute training load and use BFR as a main training modality rather than as a supplement to athletes’ normal training, and to our knowledge no previous studies have been conducted on in-season athletes (Chen, Hsieh, Ho, Ho, et al., 2022; Chen, Hsieh, Ho, Lin, et al., 2022; Giovanna et al., 2022; Held et al., 2020, 2024; Tangchaisuriya et al., 2022). Our study builds on this past work by showcasing an active BFR protocol of low intensity as a potential means to improve running time trial performance in collegiate cross country runners during their season, when used in addition to their normally prescribed running training. Yet, more work needs to be done to identify protocols and populations for which such training may have the greatest ergogenic effect.
In examination of our cardiovascular and hemodynamic outcomes, resting systolic blood pressure did not change across the study or between conditions, but resting diastolic blood pressure exceeded the smallest worthwhile change following the intervention condition and had medium effect size, with a nonsignificant trend toward lower diastolic blood pressure following the intervention condition. It is well-known that blood pressure and heart rate acutely rise to a greater degree during exercise with BFR than absolute intensity-matched exercise without BFR (de Queiros et al., 2023; Domingos & Polito, 2018; Patterson et al., 2019), whereas a recent review found that high-intensity aerobic training without BFR induced larger acute changes in heart rate andbloodpressurethanlower-intensityaerobictrainingwithBFR (de Queiros et al., 2023). When examining adaptations with BFR training,tworecentmeta-analysesincludingmainlynormotensive adults found no changes in resting blood pressure or heart rate following 6 – 12 weeks of BFR training (Maga et al., 2023; Russo et al., 2023). However, it has been documented that exercise training is more effective for reducing resting blood pressure in individuals with pre-hypertension or hypertension compared to normal blood pressure (Cornelissen & Smart, 2013), and a similar case may be true for blood pressure changes with BFR. Indeed, one study in hypertensive adults did show a reduction in both resting systolic and diastolic blood pressure following eight weeks of resistance training with BFR (Cezar et al., 2016). We
were unable to find research specifically looking at adaptations to aerobic or resistance training with BFR in athletic populations, although our findings and those from non-BFR training suggest that few changes to resting blood pressure or heart rate should be expected.
Our study also found no effect of active BFR on resting popliteal artery characteristics. This finding is in agreement with Hunt et al. (2013), who found no changes in resting popliteal artery size or function following six weeks of plantar flexion training with BFR in apparently healthy adult males. In contrast, studies by Hunt et al. (2012) and Christiansen et al. (2020) have found increases in brachial artery and femoral artery diameter, respectively, following 4 – 6 weeks of BFR training of isolated muscle groups in young adult males. Given the highly trained nature of our participants, endothelial adaptations to exercise training associated with endurance exercise (Green et al., 2017) may supersede any potential adaptations from BFR alone. Protocol differences in duration of restriction, the types of exercise completed with BFR, and the number of days/weeks may also have contributed to the mixed findings. Such possibilities should be explored in future research.
Aerobic BFR exercise causes increased physiologic strain and increased perceived exertion when compared to absolute intensity-matched exercise without restriction (Corvino et al., 2017; Ozaki et al., 2010). By contrast, our BFR protocol took place before each running workout 5+ days per week, and our participants anecdotally reported no fatigue from the protocol and noted no residual effects on their running workouts, suggesting that their self-selected light intensity walk was sufficiently light as to not hinder subsequent workout quality. Consideration of how to optimally design BFR protocols to elicit favorable training adaptations without compromising acute performance is important when planning future protocols aimed at enhancing exercise performance and recovery in rehabilitation, health, and performance settings.
Our study has several notable strengths. First and foremost, our crossover study design minimized the chances that potential effects of having different participants in control and intervention groups, or differences in training prior to beginning the study may have influenced our findings. Furthermore, our field-based time trial adds real-world value to our findings, and may highlight changes that can occur independent of hemodynamic alterations.
However, study limitations must also be acknowledged. Our sample, while of similar size to past studies in this research area, was small and relatively homogenous, and our non-significant trend toward improved running times shows that our study was underpowered to identify a potentially small but meaningful improvement in running performance with aerobic BFR. In order to control for the effects of weather, time trial course, and other potentiallyextraneousvariables,wehad torestrictourrecruitment to local cross country teams, and our rural geographic location resulted in limited ability to recruit participants meeting our strict inclusion criteria. If our study were replicated using a larger sample (possibly with high school runners or adults in running clubs), it may be sufficiently powered to detect a small benefit of aerobic BFR performed prior to training for endurance running performance. Additionally, although participants acted as their own controls, the control condition did not have a sham protocol (e.g., pumping cuffs to 20 mmHg). This choice was made to reduce burden to the athletes and to reduce risk of confusion to
participants by changing cuff pressures during the study, but it does lend the possibility that the placebo effect could have played a role inourfindings. Finally, the short washout period (2 –3days) between being in the control and intervention groups may have resulted in a residual effect on our study outcomes, although it is unlikely that it lasted three weeks until the next set of outcome measures were taken.
In conclusion, we found that daily aerobic BFR prior to run training resulted in a non-significant but potentially meaningful trend toward improved 2.4 kmtime trial performancein collegiate cross country runners without a significant change in resting cardiovascular or hemodynamic measures. More work is needed in this research area to better determine in what sports, populations, and under what BFR parameters such training strategies may elicit a chronic ergogenic effect in order to optimize their use in field-based settings.
Conflict of Interest
The authors declare no conflicts of interest.
References
Abe, T., Fujita, S., Nakajima, T., Sakamaki-Sunaga, M., Ozaki, H., Ogasawara, R., Sugaya, M., Kudo, M., Kurano, M., Yasuda, T., Sato, Y., Ohshima, H., Mukai, C., & Ishii, N. (2010). Effects of low-intensity cycle training with restricted leg blood flow on thigh musclevolume and VO2max in young men. Journal of Sports Science & Medicine, 9, 452–458. Abe, T., Kearns, C. F., & Sato, Y. (2006). Muscle size and strength are increased following walk training with restricted venous blood flow from the leg muscle, Kaatsu-walk training. Journal of Applied Physiology, 100(5), 1460–1466. https://doi.org/10.1152/japplphysiol.01267.2005
Bennett, H., & Slattery, F. (2019). Effects of blood flow restriction training on aerobic capacity and performance: A systematic review. Journal of Strength and Conditioning Research, 33(2), 572–583. https://doi.org/10.1519/JSC.0000000000002963
Buchheit, M. (2016). The numbers will love you back in return–I promise International Journal of Sports Physiology and Performance, 11(4), 551–554. https://doi.org/10.1123/IJSPP.2016-0214
Castilla-López,C.,Molina-Mula,J.,&Romero-Franco,N.(2022). Blood flow restriction during training for improving the aerobic capacity and sport performance of trained athletes: A systematic review and meta-analysis. Journal of Exercise Science and Fitness, 20(2), 190–197. https://doi.org/10.1016/j.jesf.2022.03.004
Cezar, M. A., De Sá, C. A., Corralo, V. da S., Copatti, S. L., Santos, G. A. G. dos, & Grigoletto, M. E. da S. (2016). Effects of exercise training with blood flow restriction on blood pressure in medicated hypertensive patients. Motriz: Revista de Educação Física, 22, 9–17. https://doi.org/10.1590/S19806574201600020002
Chen, Y.-T.,Hsieh, Y.-Y.,Ho,J.-Y., Ho,C.-C., Lin,T.-Y.,&Lin, J.-C. (2022). Running interval training combined with blood flow restriction increases maximal running performance and
muscular fitness in male runners. Scientific Reports, 12, 1–9. https://doi.org/10.1038/s41598-022-14253-3
Chen, Y.-T., Hsieh, Y.-Y., Ho, J.-Y., Lin, T.-Y., & Lin, J.-C. (2022). Running trainingcombined withbloodflowrestriction increases cardiopulmonary function and muscle strength in endurance athletes Journal of Strength and Conditioning Research, 36(5), 1228–1237. https://doi.org/10.1519/JSC.0000000000003938
Cohen, S. S. (1988). Statistical power analysis for the behavioral sciences (2nd Ed.). Lawrence Erlbaum Associates.
Cornelissen, V. A., & Smart, N. A. (2013). Exercise training for blood pressure: A systematic review and meta‐analysis. Journal of the American Heart Association: Cardiovascular and Cerebrovascular Disease, 2(1), 1–9 https://doi.org/10.1161/JAHA.112.004473
Corvino,R.B.,Rossiter,H.B., Loch,T.,Martins,J.C.,&Caputo, F. (2017). Physiological responses to interval endurance exercise at different levels of blood flow restriction. European Journal of Applied Physiology, 117(1), 39–52. https://doi.org/10.1007/s00421-016-3497-5
da Cunha Nascimento, D., Schoenfeld, B. J., & Prestes, J. (2020). Potential implications of blood flow restriction exercise on vascular health: A brief review. Sports Medicine, 50(1), 73–81. https://doi.org/10.1007/s40279-019-01196-5
de Lemos Muller, C. H., Farinha, J. B., Leal-Menezes, R., & Ramis, T. R. (2024). Aerobic training with blood flow restriction on muscle hypertrophy and strength: Systematic review and meta-analysis. Journal of Strength and Conditioning Research, 38(7), 1341–1349. https://doi.org/10.1519/JSC.0000000000004800
de Oliveira, M. F. M., Caputo, F., Corvino, R. B., & Denadai, B. S. (2016). Short-term low-intensity blood flow restricted interval training improves both aerobic fitness and muscle strength. Scandinavian Journal of Medicine & Science in Sports, 26(9), 1017–1025. https://doi.org/10.1111/sms.12540
de Queiros, V. S., Rolnick, N., Sabag, A., de França, I. M., Wilde, P., Vieira, J. G., Reis, V. M., Formiga, M. F., Cabral, B. G. de A. T., & Dantas, P. M. S. (2023). Acute responses in blood flow restriction low-intensity aerobic training: A metaanalysis. International Journal of Sports Medicine, 44(8), 545–557. https://doi.org/10.1055/a-2038-3635
Domingos, E., & Polito, M. D. (2018). Blood pressure response between resistance exercise with and without blood flow restriction: A systematic review and meta-analysis. Life Sciences, 209, 122–131. https://doi.org/10.1016/j.lfs.2018.08.006
Ferguson, R. A., Mitchell, E. A., Taylor, C. W., Bishop, D. J., & Christiansen, D. (2021). Blood-flow-restricted exercise: Strategies for enhancing muscle adaptation and performance in the endurance-trained athlete. Experimental Physiology, 106(4), 837–860. https://doi.org/10.1113/EP089280
Giovanna, M., Solsona, R., Sanchez, A. M. J., & Borrani, F. (2022). Effects of short-term repeated sprint training in hypoxia or with blood flow restriction on response to exercise. Journal of Physiological Anthropology, 41(1), 1–13. https://doi.org/10.1186/s40101-022-00304-1
Gnasso, A., Carallo, C., Irace, C., De Franceschi, M. S., Mattioli, P. L., Motti, C., & Cortese, C. (2001). Association between wall shear stress and flow-mediated vasodilation in healthy
Montoye et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 33-42 (2024)
men. Atherosclerosis, 156(1), 171–176. https://doi.org/10.1016/s0021-9150(00)00617-1
Green, D. J., Hopman, M. T. E., Padilla, J., Laughlin, M. H., & Thijssen, D. H. J. (2017). Vascular adaptation to exercise in humans: Role of hemodynamic stimuli. Physiological Reviews, 97(2), 495–528. https://doi.org/10.1152/physrev.00014.2016
Held, S., Behringer, M., & Donath, L. (2020). Low intensity rowing with blood flow restriction over 5 weeks increases VO2max in elite rowers: A randomized controlled trial. Journal of Science and Medicine in Sport, 23(3), 304–308. https://doi.org/10.1016/j.jsams.2019.10.002
Held, S., Rappelt, L., Rein, R., Deutsch, J.-P., Wiedenmann, T., & Donath, L. (2024). Five-week, low-intensity blood flow restriction rowing improves VO2max in elite rowers. Journal of Strength and Conditioning Research, 38(6), e299–e303. https://doi.org/10.1519/JSC.0000000000004755
Hughes, L., Paton, B., Rosenblatt, B., Gissane, C., & Patterson, S. D. (2017). Blood flow restriction training in clinical musculoskeletal rehabilitation: A systematic review and metaanalysis. British Journal of Sports Medicine, 51(13), 1003–1011. https://doi.org/10.1136/bjsports-2016-097071
Lim, C.-Y., & In, J. (2021). Considerations for crossover design in clinical study. Korean Journal of Anesthesiology, 74(4), 293–299. https://doi.org/10.4097/kja.21165
Lixandrao, M. E., Ugrinowitsch, C., Berton, R., Vechin, F. C., Conceicao, M. S., Damas, F., Libardi, C. A., & Roschel, H. (2018). Magnitude of muscle strength and mass adaptations between high-load resistance training versus low-load resistance training associated with blood-flow restriction: A systematic review and meta-analysis Sports Medicine, 48(2), 361–378. https://doi.org/10.1007/s40279-017-0795-y
Maga, M., Wachsmann-Maga, A., Batko, K., Włodarczyk, A., Kłapacz,P.,Krężel,J.,Szopa,N.,&Sliwka,A.(2023).Impact of blood-flow-restricted training on arterial functions and angiogenesis–A systematic review with meta-analysis. Biomedicines, 11(6), 1–19 https://doi.org/10.3390/biomedicines11061601
Marocolo, M., Simim, M. A. M., Bernardino, A., Monteiro, I. R., Patterson, S. D., & da Mota, G. R. (2019). Ischemic preconditioning and exercise performance: Shedding light through smallest worthwhile change. European Journal of Applied Physiology, 119(10), 2123–2149. https://doi.org/10.1007/s00421-019-04214-6
Montoye, A. H. K., McFate, D G., Cox, B. A., Rider, B. C., & Vranish, J. R. (2023). Acute effects of ischemic preconditioning at different occlusion pressures on athletic performance indicators in male soccer players. International Journal of Strength and Conditioning, 3(1). 1–10. https://doi.org/10.47206/ijsc.v3i1.221
Ozaki, H., Brechue, W. F., Sakamaki, M., Yasuda, T., Nishikawa, M., Aoki, N., Ogita, F., & Abe, T. (2010). Metabolic and cardiovascular responses to upright cycle exercise with leg blood flow reduction Journal of Sports Science & Medicine, 9(2), 224–230.
Patterson, S. D., Hughes, L., Warmington, S., Burr, J., Scott, B. R., Owens, J., Abe, T., Nielsen, J. L., Libardi, C. A., Laurentino, G., Neto, G. R., Brandner, C., Martin-Hernandez,
J., & Loenneke, J. (2019). Blood flow restriction exercise: Considerations of methodology, application, and safety Frontiers in Physiology, 10, 1–15 https://doi.org/10.3389/fphys.2019.00533
Pignanelli, C., Christiansen, D., & Burr, J. F. (2021). Blood flow restriction training and the high-performance athlete: Science to application. Journal of Applied Physiology, 130(4), 1163–1170. https://doi.org/10.1152/japplphysiol.00982.2020
Russo, A., Boppre, G., Schmidt, C., & Bohn, L. (2023). Chronic hemodynamic adaptations induced by resistance training with and without blood flow restriction in adults: A systematic review and meta-analysis. Sports Medicine and Health Science, 5(4), 259–268. https://doi.org/10.1016/j.smhs.2023.09.006
Scott, B. R., Loenneke, J. P., Slattery, K. M., & Dascombe, B. J. (2015). Exercise with blood flow restriction: An updated evidence-based approach for enhanced muscular development. Sports Medicine, 45(3), 313–325. https://doi.org/10.1007/s40279-014-0288-1
Secomb, T. W. (2016). Hemodynamics. Comprehensive Physiology, 6(2), 975–1003. https://doi.org/10.1002/cphy.c150038
Singer, J. D., & Willett, J. B. (2003). Doing data analysis with the multilevel model for change. In J. D. Singer & J. B. Willett (Eds.), Applied longitudinal data analysis: Modeling change and event occurrence (pp 75–137). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195152968.003.0004
Slysz, J., Stultz, J., & Burr, J. F. (2016). The efficacy of blood flow restricted exercise: A systematic review &meta-analysis. Journal of Science and Medicine in Sport, 19(8), 669–675. https://doi.org/10.1016/j.jsams.2015.09.005
Smith, N. D. W., Scott, B. R., Girard, O., & Peiffer, J. J. (2022). Aerobic training with blood flow restriction for endurance athletes: Potential benefits and considerations of implementation The Journal of Strength & Conditioning Research, 36(12), 3541–3550. https://doi.org/10.1519/JSC.0000000000004079
Spranger, M. D., Krishnan, A. C., Levy, P. D., O’Leary, D. S., & Smith, S. A. (2015). Blood flow restriction training and the exercise pressor reflex: A call for concern. American Journal of Physiology–Heart and Circulatory Physiology, 309(9), H1440–H1452. https://doi.org/10.1152/ajpheart.00208.2015
Tangchaisuriya, P., Chuensiri, N., Tanaka, H., & Suksom, D. (2022). Physiological adaptations to high-intensity interval training combined with blood flow restriction in masters road cyclists. Medicine and Science in Sports and Exercise, 54(5), 830–840. https://doi.org/10.1249/MSS.0000000000002857
Vranish, J. R., Young, B. E., Kaur, J., Patik, J. C., Padilla, J., & Fadel, P. J. (2017). Influence of sex on microvascular and macrovascular responses to prolonged sitting. American Journal of Physiology. Heart and Circulatory Physiology, 312(4), H800–H805. https://doi.org/10.1152/ajpheart.00823.2016
Wortman, R. J., Brown, S. M., Savage-Elliott, I., Finley, Z. J., & Mulcahey, M. K. (2021). Blood flow restriction training for athletes: A systematic review TheAmericanJournalofSports Medicine, 49(7), 1938–1944. https://doi.org/10.1177/0363546520964454
The Journal of Sport and Exercise Science, Vol. 8, Issue 1, 43-64 (2024)
www.jses.net
2703-240X
An expected wins approach using Fisher’s Exact Test to identify the bogey effect in sports: An application to tennis
Rory P. Bunker1* , Calvin Yeung1 , Keisuke Fujii1,2,3
1Graduate School of Informatics, Nagoya University, Japan
2RIKEN Advanced Intelligence Project, Japan
3PRESTO Japan Science and Technology Agency, Japan
Received: 31.10.2023
Accepted: 05.06.2024
Online: 22.11.2024
Keywords:
Tennis
Bogey
Betting odds
Fisher’ s Exact Test
Elo ratings
In sports, so-called “bogey” players or teams tend to beat a particular opposition with regularity despite being of similarability. Although the existence of bogey playersis widely discussed and debated among sports fans and the media, methods that could be used to identify the bogey effect have received little attention in the literature. This study proposes a statistical procedure to identify bogey players using a publicly available men’s AssociationofTennisProfessionals(ATP) andWomen'sTennisAssociation (WTA)dataset, which is also split into Grand Slam and non-Grand Slam matches. The proposed method iterates over all unique player pairs and applies Fisher’s Exact Test to a contingency table containing expected wins and actual win distributions for historical matches between the players in the player pair. To compute expected wins, betting odds- or Elo ratings-implied probabilities for each player are aggregated over all matches between the player pair for each player. If the Fisher’s Exact Test result is statistically significant (that is, actual wins and expected wins do not follow the same distribution), and the expected wins and actual win counts are contradictory, we suggest that the bogey effect exists between the two players. The obtained results suggest that the bogey player effect exists in professional tennis but is rare (and even rarer in Grand Slams), and expected wins obtained using betting odds-implied win probabilities more closely matched actual wins than Elo ratings, which resulted in fewer bogey player pairs being identified with betting odds. The number of bogey player pairs identified is intuitively found to be inversely related to the predictability of matches.
1. Introduction
The existence of the bogey effect, in which players (or teams in the case of team sports) tend to beat a particular opposition with regularity despite being of similar ability, is widely discussed amongsportsfansandthemedia.Forinstance,Franceatonepoint was considered the bogey team of New Zealand in Rugby World Cup tournaments (Bruce, 2014), while in soccer, Italy was considered the bogey team of Germany (Wilms, 2013). While tennis, which is considered in the current study, appears to have had less attention in the media than soccer, there has been some
discussion of potential bogey players in online forums and in articles (Niall, 2013; Wood, 2017). Loosely speaking, a bogey player/team, or “Angstgegener” (translated as “feared opponent”) as it is known in the German language, tends to habitually beat another specified player/team despite appearing to be of equal, or even lesser, strength on paper. Although the bogey phenomenon has been mentioned in a small number of academic studies in, for example, education (Bruce, 2014) and sociology (Chiweshe,2021; Poulton, 2004), and in doctoral theses (Awerbuch, 2009; Wilms, 2014), it has been largely unexplored in the sports science and sports statistics disciplines.
*Corresponding Author: Rory P. Bunker, Graduate School of Informatics, Nagoya University, Japan, rorybunker@gmail.com
Relatedtothebogeyeffectaretheconceptsofstreaks,form,hot hand,stability(non-stationarity),autocorrelation,and“hotand cold nights”. Streaks can be considered both in terms of individual player actions (e.g., home runs in baseball, three-pointers in basketball)ormatch-winning streaks.Form,alsoknownasthe “hot hand”phenomenoninbasketball,assumesthatfutureoutcomescan be determined, at least partly, based on the most recent outcomes and that players or teams having successful streaks impact their futuresuccesses(Ayton&Fischer,2004;Bar-Eli,Avugos,&Raab, 2006). Carlson and Shu (2007) found that across five diverse studies, including one related to shooting in basketball, the third repeated event within a sequence is critical to the subjective belief that a streak is happening. The most common techniques used in such analyses have been Wald-Wolfowitz Runs Tests and autocorrelation tests (Carlson & Shu, 2007; Peel & Clauset, 2015; Raab,2012;Stone,2012).Hales(1999)arguedthatautocorrelation, the correlation between the outcomes of consecutive events, and non-stationarity, the probability of success fluctuating over time, shouldbeconsidered separately becausethetwoconceptsrepresent different underlying mechanisms of the hot hand effect. Specifically, (positive) autocorrelation, which is often measured using the correlation coefficient or the runs test, suggests that success in one event increases the likelihood of success in a subsequent event, thus indicating a hot hand effect. If autocorrelation is present, it may indicate that performance is influenced by recent success/failure, and a measure of positive association between shot outcomes may be appropriate. Nonstationarity may be caused by several factors (e.g., form, fatigue, or external factors such as opponent performance) and the chi-square test can be used to detect changes in the probability of success over time. Non-stationarity indicates the existence of fluctuations in player performance, and a time-varying ability parameter (e.g., time-varying Bradley-Terry parameters as per Cattelan, Varin, & Firth, 2013) may be appropriate to model the data. Steeger, Dulin, andGonzalez(2021)distinguishedbetweenstreaksandmomentum; the former referring to observed sequences of events each of which may or may not have dependence between them, while momentum suggests that a dependence exists between events that are similar. In their seminal paper, Gilovich, Vallone, and Tversky (1985) refer to basketball shooters having “hot” and “cold” nights (i.e., strong andpoorperformance,respectively),andanalysedwhetherstability exists across matches in terms of shooters having more hot or cold nightsthanwouldbeexpectedbychance;thatis,howthevariability in match shooting percentages that is observed compares with the expected variability according to a player’s record overall. The “gambler’s fallacy”, also known more generally as negative recency, is the belief that in sequences comprised of binary random events, runs of a specific outcome will be balanced by corrective action; that is, a tendency for the other outcome (Estes, 1964 and Ayton & Fisher, 2004, as cited in Steeger, Dulin, & Gonzalez, 2021). Positive recency, also known as the hot-hand fallacy, is the inclination to predict future outcomes the same as recent outcomes (Ayton & Fischer, 2004, as cited in Steeger, Dulin, & Gonzalez, 2021). Baboota and Kaur (2019) engineered streak and timeweighted streak features for soccer match result prediction that considertheresultsofasingle, andaformfeaturethatconsidersthe results between specific pairs of teams. At first glance, it can be temptingtoattribute along streakofwins toaparticular team tothe bogey effect. Tottenham, for example, won only one game out of 37 against Chelsea between 1990 and 2006, and Watford did not
beat Manchester City in the 30-year period from 1989 to 2019. However, it may have been the case that these results may all have been expected; thus, the question then becomes how expectedness can be accounted for. The media tend to use the “bogey” term relatively loosely, without providing additional contextual information that would help determine whether such results are actually unexpected. This study attempts to clarify this by introducing a bogey player identification method that uses Fisher’s Exact Test to determine whether the match results between a particular pair of tennis players deviate from what would be expected, given the betting odds or Elo ratings of the two players in each player pair.
Prior studies that have focused on bogey effect identification specifically have attempted to use statistical techniques including the Wald-Wolfowitz runs test (Bunker, 2022) and the Aylmer test (Hankin & Bunker, 2016), using data from Tennis and Rugby League. Using runs tests, however, also tended to identify result sequences evenly split between unexpected wins and unexpected losses, which is not indicative of the bogey effect, and the Aylmer test, since it was applied to the match results of all pairs simultaneously, was subject to the multiple comparisons issue. In the current study, betting odds, which were used to identify unexpected results in Bunker (2022), as well as Elo ratings are used to compute expected wins for each player pair. In particular, (implied) win probabilities for both betting odds and Elo ratings are aggregated to calculate expected wins for each player in the player pair. Leitner, Zeileis, and Hornik (2009) proposed a bookmaker consensus model that aggregates bookmaker expectations into a prediction for tennis match results. In the proposed method, the odds that we use in the dataset already represent the average across multiple bookmakers, and we aggregate the odds-implied win probabilities of each player in the player pair. Fisher’s Exact Test (FET) is then applied to a contingency table consisting of the computed expected wins distribution and the actual win distribution for the player pair.The bogey effect is considered to exist between two players if FET yields a statistically significant result, and the expected wins and actual wins contradict. The method proposed circumvents the multiple comparisons problem since it is applied iteratively to each unique player pair, and match results and betting odds are independent since odds are determined by bookmakers prior to matches and thus represent their assessment of the probabilities of the match outcomes, and bets placed on certain outcomes do not affect match results, while match result is determined purely by the teams' or players' in-match performance. While odds can provide insights into bookmaker expectations, in the absence of match-fixing, they do not have any effect on match results. As well as proposing a novel method for bogey effect identification, the method is demonstrated on publicly available datasets consisting of 33,976 men’s Association of Tennis Professionals (ATP) matches from 2005 to 2020 and 27,094 Women's Tennis Association(WTA) matchesfrom2007to2020,aswellassubsets of the original datasets comprising only Grand Slam and nonGrand Slam matches. The analysis presented in this paper is also relevant in the context of sports psychology in that some players (teams) struggling more against a particular opposition player (team) may be a psychological phenomenon.
We hypothesise that the betting odds are more accurate for Grand Slam matches since, first, bookmakers would carry out more research and perform more modelling before setting the
odds for the match outcome, and second, the betting volume is likely to be higher on Grand Slam matches compared to nonGrand Slam matches, and this additional information from betting volume increases the accuracy of the odds. Since betting odds accuracy is inversely related to the number of bogey players identified by odds in our proposed method, both of these factors would contribute to fewer bogey players being identified.
On the other hand, Elo ratings are affected only by match results and are updated dynamically over time based on these match outcomes. Form and strength would also be accounted for bybookmakers,andthey mayeven useEloratingsorotherratings in their models to set odds, however, this is only a subset of the information they would use to set betting odds.
Through a regression-based analysis of 51,881 tennis matches, Barrutiabengoa, Corredor, and Muga (2022) found that, even after controlling for surprise factor uncertainty and the amount of media attention, the prices that bookmakers quote are higher for women’s matches than men’s, which suggests that the betting volume (and, therefore, betting odds accuracy) on women's matches could be lower compared to men's. Vaughan Williams, Liu, Dixon, and Gerrard(2021)comparedtheperformanceofbettingodds,rankings, standard and surface-specific Elo ratings, and weighted rating composites, including and excluding the betting odds, in predicting men’s and women’s professional tennis matches and found that betting odds performed well in general, and standard Elo ratings performed well for women’s tennis. The authors found that Elo and bettingoddsperformedbetterthanrankings,whichsupportstheuse of these two variables in our proposed method.
Kovalchik (2016) found that the accuracy of predictive models when predicting match results is markedly different for lowerranked and top-ranked players, finding that match outcome predictionmodelsare10%to20%lessaccurateformatchesamong lower-rankedplayersthanmatchesamongtop-rankedplayers.Yue, Chou, Hsieh, and Hsiao (2022) showed that win probability with respect to ranking difference fluctuates to a greater degree (i.e., predictability decreases) when the ranking difference increases (Figure 2, Yue et al., 2022) due to smaller number of samples at larger ranking differences (Figure 3, Yue et al., 2022). The findings of Yue et al. (2022) imply then that matches between top-ranked players have a small ranking difference (Elo rating difference) and are thus easier to predict, while matches between top-ranked and lower-ranked players, which have a large ranking (Elo rating) difference, are more difficult to predict. In this study, we partition the original dataset into Grand Slam and non-Grand Slam matches. Grand Slams generally consist of matches among top-ranked players, thus, based on the findings of Kovalchik (2016) and Yue et al. (2022), we would expect our proposed method to identify fewer bogey player pairs for the Grand Slam match dataset compared to the non-Grand Slam match dataset.
Elo ratings, which are updated over time based on historical match results, do not incorporate factors such as the court surface and what hand the players play with, while betting odds do incorporate such factors through the odds set by bookmakers supplemented by fan knowledge reflected in betting volume, which bookmakers use to tweak their initial odds. We would
expect that using Elo ratings in the proposed method will generally identify more bogey players than betting odds.
2. Methods
2.1. Data
2.1.1.
Datasets
Publicly available data from professional ATP (men’s) and WTA (women’s) tennis was sourced for this study. In particular, we use the same datasets used by Angelini, Candila, and De Angelis (2022), which were originally sourced from the website tennisdata.co.uk. Among many other variables, the dataset contains match data from ATP and WTA tournaments and Grand Slams, as well as bookmaker odds, player rankings, and ATP/WTA player points. The datasets contain 33,976 men’s ATP matches from 2005to 2020 and27,094women’s WTA matches from2007 to 2018. The ATP and WTA datasets were downloaded as RData files from “Appendix C. Supplementary materials” in the paper by Angelini, Candila, and De Angelis (2022). An R script was then created to clean the data using the clean() function in the welo R package (Candila, 2023) and to convert and export the cleaned data to a CSV file. The datasets were passed through the welo R package’s clean() function, which reduced the number of matches in the final dataset. This data-cleaning function reduced theoriginal numberofmatches from 38,868 to 33,976 for the ATP dataset and from 30,706 to 27,094 for the WTA dataset.1
For further analysis, the ATP and WTA datasets consisting of all matches were each divided into two additional datasets consisting of Grand Slam matches and non-Grand Slam matches (Figure 1), and the proposed method will also be applied to these four additional datasets.
Figure 1: Datasets used upon which the proposed method is applied.
1 The clean() function performs ten cleaning operations, which are described under the clean function details on page 4 in the welo package manual: https://cran.r-project.org/package=welo. The default options of the function were used.
2.1.2. Descriptive statistics
To compare the temporal variability of betting odds (winner, loser, and combined) and Elo ratings, we plot the coefficient of variation (CV), which is calculated by dividing the standard deviation by the meanandisindicativeofthelevelofdispersionaroundthemeanand accountsforthedifferentscaleofvariance/standarddeviationofodds and Elo ratings, of each for both the ATP and WTA datasets (Figure 2A and Figure 3A). As can be seen, the variability of Elo ratings is much lower overall compared to betting odds. It is notable from Figure 3 that the variability of betting odds for women’s tennis declined over the time period 2005 and 2020. The variability of Elo ratingsdeclinedforbothATPandWTAinthelatterperiod,however, WTA Elo rating variability declined from 2013 onwards, whereas
Figure 2: Coefficient of Variation (CV) of Betting Odds and Elo Ratings for each year (A), as well as the more granular view of the mean and CV for each year of Elo rating and Betting Odds (B and C, respectively) for the ATP dataset.
ATP Elo rating variability declined later from 2016 onwards. The mean Elo rating generally increased over the time period for both ATP and WTA. Another noticeable feature of Figures 2C and 3C is that the mean betting odds and betting odds CV for the WTA, especiallyfortheloserodds,declinedoverthe sampleperiod. Onthe other hand, the mean betting odds and betting odds CV for the ATP had no discernible trend over the sample period. This perhaps indicates that bookmakers, on account of an evening in the level of competition in the WTA, as evidenced by the decline in Elo rating CV from 2013 onwards, began to lower the price on the likely losers inWTAmatches.TheremainingfiguresfortheGrandSlamandnonGrand Slam datasets for the ATP and WTA are in the Supplemental material (Supplementary Figures 1 to 4)
Figure 3 Coefficient of Variation (CV) of Betting Odds andElo Ratings for each year (A), as well as the more granular view of the mean and CV for each year of Elo rating and Betting Odds (B and C, respectively) for the WTA dataset.
(A)
(A)
(B)
(B)
(C)
(C)
2.2. Proposed method
The proposed method consists of three steps, which are depicted in Figure 4 and are outlined in the following three subsections.
Figure 4: Three steps of the proposed method.
2.2.1. Computing expected wins
As mentioned, expected wins are determined using two approaches: using betting odds- and Elo ratings-implied probabilities. So that the win probabilities for each player in each match add to one, the betting odds-implied win probabilities are obtained by dividing the reciprocal of the decimal betting odds by a normalization factor (also known as the over-round), which is simply the sum of the two decimal betting odd reciprocals. Mathematically, the win probabilities for each player, �� and ��, in match ��, can thus be derived from the decimal betting odds by:
Similarly, the Elo-implied estimated win probability for player �� over player �� in match �� is given by:
and
where ���� �� and ���� �� arethedecimalbettingoddsforaplayer �� winand player �� win, respectively, in match ��. The denominator in each of these two expressions is the normalization factor (over-round).
The second way expected wins are determined is based on Elo ratings-implied win probabilities. It can be shown that the Bradley-Terry strength/ability parameter can be expressed as a function of the Elo rating (Coulom, 2007). In the Bradley-Terry model, the probability that �� beats �� is given by:
where the Elo rating of
is defined
or equivalently
Therefore, the Elo-implied estimated win probability for player �� over player �� in match �� is given by:
Note that, unlike betting odds, normalization is not required as it is already the case with Elo-rating estimated probabilities that ���� �� + ���� �� =1. The before- and after-match Elo ratings were computed by passing the cleaned ATP and WTA datasets (see 2.1. Data) to the welofit() function in the welo package (the default options were used). For estimating win probabilities, we use the before-match Elo ratings. Since the betting odds and Elo ratings already account for historical information, it is reasonable to ignore the temporal order of the match result data (i.e., we do not need to consider the match results as a temporally ordered sequence).
2.2.2. Fisher’ s Exact Test
A Fisher’s Exact Test is applied to compare each pair of players’ expected wins distribution with their actual matchresult distribution. The bogey effect is assumed to represent a violation of expectation. The obtained expected wins distribution for each player in each player pair comprises one part of the contingency table to which the Fisher’sExactTest,whichhastheadvantageofbeingabletobeused for small sample sizes, is applied. The expected wins for each player in a given player pair are obtained by simply summing the estimated win probabilities across all matches they have played against that specific player, which represents the expected number of matches eachplayershouldwingiventhebettingodds/Eloratingsdistribution. The other part of the contingency table is the actual win distribution between the two players. The null hypothesis, H0, and alternative hypothesis, H1, are described as follows:
H0: The expected and actual win distributions are the same, or there is no significant difference between them.
H1: The expected and actual win distributions are not the same, or there is a significant difference between them.
The contingency table for the Fisher exact test for one pair of players, player �� and player ��, is shown in Table 1. In Table 1, �� represents the total number of historical matches played between player �� andplayer ��.Thus, ∑ ���� �� �� ��=1 representstheaggregationofthe betting odds- or Elo-implied win probabilities for player �� over all historicalmatchesagainstplayer �� Similarly, ∑ ���� �� �� ��=1 representsthe aggregation of the betting odds (or Elo) implied win probabilities for player �� over all historical matches against player ��. The cell ���� �� in Table 1 is a binary variable that takes the value of 1 if player �� won against player �� in match �� and 0 otherwise, and analogously, ���� �� is a binary variable that takes the value of 1 if player �� beat player �� in match �� and 0 otherwise. Therefore, ∑ ���� �� �� ��=1 represents the total number of actual wins player �� has had over player �� in their �� past matches and ∑ ���� �� �� ��=1 the total number of actual wins player �� has had over player �� in their �� past matches.
Table 1: Contingency table for Fisher’s Exact Test for a given player pair over their N historical matches.
Betting Odds- or Elo ratingsimplied win probabilities Actual match result
Player
Player
2.3. Bogey effect identification
To identify the bogey effect between a pair of players, the method iteratively computes the Fisher’s Exact Test p-values for each player pair. Having a p ≤ ��, where �� is a specific significance level (in this study we consider �� = 0.05 and �� = 0.1, which means that we have 95%or90%confidencethatabogeyplayerpairexists),isanecessary but not sufficient condition for the bogey effect to exist. The p-value is calculated from the Fisher’s Exact Test that is applied to the contingency table structure for the �� historical matches between a particular player pair as per Table 1. If Fisher’s Exact Test yields a statistically significant result and the expected wins and actual wins contradict thatis,playerA wasexpected towinmorematches than player B but actually player A won fewer, or player B was expected to win more matches than player A but actually player B won fewer we suggest that the bogey effect exists between the two players. Using the notation in Table 1, we suggest that the bogey effect exists between apair ofplayersplayer �� andplayer ��,based on their �� past matches, if the following holds true:
Table 2: Summary of the number of significant player pairs with the bogey effect identified for each of the six datasets, using aggregated betting odds- and Elo ratings-implied probabilities for computing expected wins, at the 95% and 90% significance levels.
95% level of significance 90% level of significance Player pairs (n) Betting Odds Elo Ratings Betting Odds Elo
Non-Grand Slam
Notes: The player pairs that are significant at the 95% significance level are also significant at the 90% significance level. For example, for ATP data with Elo ratings, the 18 significant bogey player pairs at the 90% level of significance also include the 3 player pairs that are significant at the 95% level.
Since the total number of player pairs differs across datasets (see the third column in Table 2), in order to make a like-for-like comparison between datasets, the number of bogey player pairs identified is scaled by the total number of player pairs in the dataset in Figure 5. In particular, Figure 5 shows the number of bogey effect player pairs identified for each dataset using betting odds- and Elo ratings-implied probabilities, as a percentage of the total number of player pairs in each dataset, at the 95% (Figure 5A) and 90% (Figure 5B) significance levels.
3. Results
3.1. Number of bogey player pairs identified with the proposed method using Elo ratings and betting odds for each dataset
A summary of the results for all datasets, at the 90% and 95% level ofconfidence,isshowninTable2.Aninitialobservationfrom Table 2 is that, regardless of the level of significance used and whether bettingoddsorEloratingsareused,thenumberofbogeyplayerpairs identified is very small relative to the number of player pairs in the datasets. This suggests that the bogey player effect is very rare in professional tennis.
For the whole ATP dataset, of the 18,241 distinct ATP player pairs, 4 and 18 significant bogey pairs were identified at the 90% significance level using betting odds and Elo ratings, respectively. ForthewholeWTAdataset,ofthe15,844distinctWTAplayerpairs, 7 and 13 significant bogey pairs were identified using betting odds and Elo ratings, respectively, at the 90% level of significance.
Some observations can be made from Figure 5 and Table 2. The proposed method with betting odds-implied probabilities obtained fewer bogey player pairs (as a percentage of total player pairs) for all datasets apart from the ATP Grand Slam match dataset. Across the ATP Grand Slam and WTA Grand Slam datasets with both Elo ratings and betting odds-implied probabilities used to compute expected wins, and the ATP Grand Slams dataset with betting odds-implied probabilities used, only one bogey player pair was identified. These results lend support to what was expected: that Grand Slams are closely followed by bookmakers and thus odds are set accurately, and the small differences in Elo ratings between matches among generally topranked players at Grand Slam tournaments result in fewer bogey player pairs being identified compared to in the non-Grand Slam match dataset. While the proposed method using betting odds identified relatively fewer bogey player pairs in men’s than women’s tennis in the all-match datasets, this wasn’t the case with the Grand Slam or non-Grand Slam subsets. Somewhat consistent with Vaughan Williams et al. (2021), Elo ratings appear to perform well in the WTA since there were relatively fewer bogey player pairs identified in the WTA using Elo ratings compared to the ATP.
Figure 5: The number of bogey effect player pairs identified for each dataset using betting odds- and Elo ratings-implied probabilities, as a percentage of the total number of player pairs in each dataset, at the 95% (A) and 90% (B) significance levels (data from Table 2).
3.2. Visualisingobtainedbogeyplayerpairsforaparticulardataset
Since the order of the players in each player pair is not material, we visualise bogey player pairs on the absolute difference in actual wins against the absolute difference in expected wins. The size of these points is scaled based on the FET p-value, and significant playerpairs basedonthe FET thatarenot bogeyplayerpairs canbe distinguished by shape. Figures 6 to Figure 9 depict the significant bogey (and significant non-bogey) player pairs in this manner for the all-match ATP and WTA datasets, obtained with Elo ratingsand betting odds-implied probabilities (Supplementary Figures 5 to 8; Supplementary Tables 6 to 9 show the same plots and corresponding data for the Grand Slam and non-Grand Slam datasets for the ATP and WTA). These figures correspond to the data in Supplementary Tables 1, 2, 3, and 4 in the Supplemental material
In general, Figures 6 to Figure 9 exhibit a cluster of points towards the bottom-left hand corner with an absolute actual win difference between the players in the bogey player pair of around two and an absolute expected wins difference of around three. There is often an additional cluster of points with roughly the same absolute expected wins difference but higher absolute actual wins difference values. In all figures there was also an outlier player pair withhigherabsoluteactualandexpectedwinsdifferences,however, these were generally significant but not identified as a bogey player pair since their actual and expected wins did not contradict.
Figure 6: Visualizing, on axes representing the absolute difference in actual wins and expected wins for the player pair, the statistically significant (with at least 90% confidence) and bogey player pairs from theATPtennisdataset,withaggregatedEloratings-impliedprobabilities used to compute expected wins. The data used to create this plot is presentedinSupplementaryTable1intheSupplementalmaterial
Figure 7: Visualizing, on axes representing the absolute difference in actualwinsandexpectedwinsfortheplayerpair,statisticallysignificant (with at least 90% confidence) and bogey player pairs from the ATP tennisdataset,withaggregatedbettingodds-impliedprobabilitiesusedto computeexpected wins.Thedataused to create this plotis presented in SupplementaryTable2intheSupplementalmaterial
Figure 8: Visualizing, on axes representing the absolute difference in actualwinsandexpectedwinsfortheplayerpair,statisticallysignificant (withatleast90%confidence)bogeyplayerpairsfromtheWTAtennis dataset, with aggregated Elo ratings-implied probabilities used to computeexpected wins.Thedataused to create this plotis presented in SupplementaryTable3intheSupplementalmaterial.
Figure 9: Visualizing, on axes representing the absolute difference in actualwinsandexpectedwinsfortheplayerpair,statisticallysignificant (withatleast90%confidence)bogeyplayerpairsfromtheWTAtennis dataset, with aggregated Betting odds-implied probabilities used to computeexpected wins.Thedataused to create this plotis presented in SupplementaryTable4intheSupplementalmaterial
3.3. Analysing the overlap in the bogey player pairs identified by Elo ratings and betting odds
Figure 10A shows the number of bogey player pairs identified by bettingoddsbasedonwhethertheywerealsoidentifiedbyEloratings (oronlybettingodds).Figure10Bshowsthenumberofbogeyplayer pairsidentifiedbyEloratingsbasedonwhethertheplayerpairswere also identified by betting odds (or only by Elo ratings).
Figure10:Thenumberofbogeyplayerpairsidentifiedby(A)betting odds, split into whether they were also identified by Elo ratings or onlybybettingodds,and(B)Eloratings,splitintowhethertheywere also identified by betting odds or only Elo ratings.
For example, for the whole ATP dataset, of the significant bogey player pairs identified by Elo ratings (Supplementary Table 1), three of these (Ramirez-Hidalgo R. vs Verdasco F.; Kohlschreiber P. vs Mayer L.; and Benneteau J. vs Gulbis E.) were also identified by bettingodds(SupplementaryTable2).Fifteenbogeyplayerpairsthat were identified by Elo ratings were only identified by Elo but were not identified by betting odds.
Figure 10A suggests that the bogey player pairs identified by the proposed method with betting odds were largely also identified with Eloratings.However,Figure10Bsuggeststhatthebogeyplayerpairs identified by the proposed method with Elo ratings were largely not also identified with betting odds (the WTA all-match dataset is one notable exception).
3.4. Visualising the expected win distribution violation quantification foreachdataset
Figure 11 shows average expected and actual win probability, as well as the differences, for player pairs containing and not containingbogeyplayers.Theunderlyingdataforthisplotisshown in Supplementary Table 12 in the Supplemental material. For each player pair type, whether the player pair is a bogey player pair or not, the average expected wins and average actual wins were calculatedbyscalingtheactualandexpectedwinsbasedonthetotal number of matches between the players in the player pair. Taking the difference between these two values provides a means of quantifying the degree to which the expected win distribution is violated. The red values in Figure 11 denote the average difference in expected and actual win probabilities for player pairs without a bogey player, while the blue values denote the average difference in expected and actual win probabilities for player pairs that don’t involve a bogey player. Three of the datasets/methods on the farright have no bogey player pairs (see Table 2) and therefore only two points show for these. As we would expect based on our proposed method’s design, for player pairs that do contain a bogey player, there is a large violation of the expected wins distribution in terms of the average difference in expected and actual wins (the blue values), and are many times larger than the violation of the expected distribution for player pairs not containing a bogey player (red values). It is also notable that while the values representing the violation of the expected wins distribution for bogey player pairs wereallrelativelysimilar,rangingfrom0.665to0.708,whereasthe values representing the violation of the expected wins distribution values for bogey player pairs while smaller in magnitude ranged from 0.0389 to -0.0382.
4. Discussion
This study proposed a bogey player identification method that involves computing an expected wins distribution using the summed implied win probabilities based on Elo ratings and betting odds, constructing a contingency table containing actual and expected wins between each player in a player pair, and applying Fisher’s Exact Test to this contingency table. If a significant result from Fisher’s Exact Test was obtained, and the actual wins and expected wins contradict, the bogey effect was deemed to exist between the two players in the player pair.
The obtained results suggest that although the bogey player effect exists in professional tennis, it is very rare. The proposed
(A)
(B)
Figure 11: For each dataset and method (Elo/odds), this plot shows the average expected wins and average actual wins scaled based on the number of matches between players in a player pair, as well as their differences for player pairs that contain and do not contain bogey players (at the 90% level of confidence) (see Supplementary Table 12 in the Supplemental material for the underlying data in this plot).
methodwithbettingoddsusedtocomputeexpectedwinsobtained fewer bogey player pairs, as a percentage of total player pairs in the dataset, for all datasets except for the ATP Grand Slam dataset. Onlyonesignificant(atthe90%levelofconfidence)bogeyplayer pair was identified across the four Grand Slam datasets: ATP and WTAGrandSlamdatasetsusingEloratingsandbettingodds.The resultsconformed with ourprior expectation that Grand Slams are closely analysed by bookmakers and so odds are set accurately, which results in fewer bogey player pairs in general. Furthermore, when the proposed method is used with Elo ratings to compute expected wins, the small differences in Elo ratings in matches among generally top-ranked players at Grand Slams mean that fewer bogey player pairs are identified compared to non-Grand Slams. Although betting odds identified relatively fewer bogey player pairs in ATP than WTA tennis in the ATP and WTA datasets as a whole, this did not hold for the Grand Slam or nonGrand Slam datasets. Since there were relatively fewer bogey player pairs identified in the WTA using Elo ratings compared to the ATP, Elo ratings could be said to be of better predictive value for the WTA than the ATP, a result that is consistent with Vaughan Williams et al. (2021). Surprisingly, when visualising obtained FET-significant player pairs and bogey player pairs for the various datasets, certain patterns emerged in terms of the clusters of bogey player pairs’ absolute differences in actual and expected wins. However, these patterns/clusters did not appear useful for identifying which of the pairs are actually bogey player pairs. When analysing the overlap in the bogey player pairs that were identifiedby Eloratingsandbettingodds,whilebettingodds
generally identified fewer bogey player pairs than Elo ratings, the majority of the bogey player pairs identified by betting odds were also identified by Elo ratings. On the other hand, the majority of bogey player pairs identified by Elo ratings were only identified by Elo ratings but not by betting odds. This suggests perhaps that bettingoddsmaybegenerallyamorereliablemeansofcomputing expected wins and thus identifying bogey player pairs. When visualising the expected win distribution violation quantification for the various datasets by considering the average difference in expected and actual wins for player pairs containing and not containing bogey players, we validated that, for bogey player pairs, there was a large violation of the expected wins distribution in terms of the average difference in expected and actual wins, and these differences – which were relatively similar across the different datasets/methods (Elo and odds) – were many times larger than the violation of the expected distribution for nonbogey player pairs.
Analysing a particular player’s performance, whether they are pronetobeinga bogey playeror being thebogey playerofanother, can be useful for player-level performance analysis and match preparation. For instance, Stosur is a WTA player who appeared both as a bogey player and as a non-bogey player in bogey player pairs, which is interesting from a practical performance analysis perspective, for example, for her opponents and coaching staff to analyse further.
The proposed method is flexible in that it can be applied not only to tennis but to other sports, directly to sports with two outcomes and with some modifications to sports with more than
two outcomes. In future work, the method could therefore be applied to other sports with two outcomes (e.g., basketball), and it could be extended to sports with three outcomes, for example, soccer, by using extensions of Fisher’s Exact Test that can handle contingency tables with more than two columns/rows, for example, the Freeman-Halton extension of Fisher’s Exact Test (Freeman & Halton, 1951). For instance, to include a draw outcome, in addition to summing the win probability for each opposition as in the current study, the probability of a draw multiplied by 0.5 for both players would be summed, and 0.5 could be used to represent an actual draw result. Rating systems other than Elo ratings could also be trialled. Finally, other subsets oftheoriginaldatasetotherthan GrandSlam/non-GrandSlam,for example, based on court surface, time period, player hand, and player rank group could be considered. For example, based on Figure 3 above, splitting the WTA dataset into 2005 to 2012 and 2013 to2020 would be an obvious partitionoftheoriginaldataset. Also, subsets of the original dataset based on Elo rating differences would also be interesting to investigate. For instance, since bookmakers have only a limited (or no match) history to go on in the case of matches among two low-ranked players, odds may be more difficult to set, and therefore their value for prediction may be lower and thus more bogey player pairs would beidentified.Matchesamonglow-rankedplayerswould,however, have a small ranking difference in terms of Elo ratings and, therefore, based on the findings of Yue et al. (2022) would be more predictable and we could hypothesise that this would result in fewer bogey player pairs compared to when odds are used.
Conflict of Interest
The authors declare no conflict of interests
Acknowledgment
This work was supported by JSPS under Grant [number 20H04075] and JST Presto under Grant [number JPMJPR20CA].
Data availability
The dataset that supports the findings of this study was obtained, asdescribed inthe“Data” subsection of“MaterialsandMethods”, from the openly available data in the Appendix of the online version of the paper by Angelini, Candila, and De Angelis (2022), at doi:10.1016/j.ejor.2021.04.011 under “Supplementary Data S1”. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/
Code
The code is available at the following GitHub repository: https://github.com/rorybunker/bogey-phenomenon-sport/
References
Angelini, G., Candila, V., & De Angelis, L. (2022). Weighted Elo rating for tennis match predictions. European Journal of Operational Research, 297(1), 120–132.
Ayton, P., & Fischer, I. (2004). The hot hand fallacy and the gambler’s fallacy: Two faces of subjective randomness? Memory & Cognition, 32(6), 1369–1378.
Awerbuch, W. (2009). Anwendung von Data Mining zu statistischen Auswertungen und Vorhersagen im Sport [Master’s thesis, TU Darmstadt]. TU Darmstadt Knowledge Engineering website. https://ketud.github.io/lehre/arbeiten/diplom/2009/Awerbuch_Wladim ir.pdf
Baboota, R., & Kaur, H. (2019). Predictive analysis and modelling football results using machine learning approach for English Premier League. International Journal of Forecasting, 35(2), 741–755.
Barrutiabengoa, J. M., Corredor, P., & Muga, L. (2022). Does the betting industry price gender? Evidence from professional tennis. Journal of Sports Economics, 23(7), 881–906.
Bar-Eli, M., Avugos, S., & Raab, M. (2006). Twenty years of“hot hand” research: Review and critique. Psychology of Sport and Exercise, 7(6), 525–553.
Bunker, R. (2022). The Bogey Phenomenon in Sport. In J. J. Reade (Eds.), IX Mathsport International 2022 Proceedings (pp.15–21).
https://www.mathsportinternational.com/MathSport2022Proc eedings.pdf
Candila, V. (2023). welo: An R package for weighted and standard Elo rates. Statistica Applicata-Italian Journal of Applied Statistics, (1)
Carlson, K. A., & Shu, S. B. (2007). The rule of three: How the third event signals the emergence of a streak. Organizational Behavior and Human Decision Processes, 104(1), 113–121.
Chiweshe, M. K. (2021). Frenemies: Understanding the interconnectedness of rival fan identities in Harare, Zimbabwe. In K. Bandyopadhyay (Eds.), Face to Face: Enduring Rivalries in World Soccer (pp. 191–203). Routledge.
Cattelan, M., Varin, C., & Firth, D. (2013). Dynamic Bradley–Terry modelling of sports tournaments. Journal of the Royal Statistical SocietySeriesC: Applied Statistics, 62(1), 135–150. Coulom, R. (2007). Computing “elo ratings” of move patterns in the game of go. ICGA Journal, 30(4), 198–208.
Estes, W. K. (1964). Probability learning. In A. Melton (Ed.), Categories of human learning (pp. 89–128). Academic Press/Elsevier
Freeman, G. H., & Halton, J. H. (1951). Note on an exact treatment of contingency, goodness of fit and other problems of significance. Biometrika, 38(1/2), 141–149.
Gilovich, T., Vallone, R., & Tversky, A. (1985). The hot hand in basketball: On the misperception of random sequences. Cognitive Psychology, 17(3), 295–314.
Hales, S. D. (1999). An epistemologist looks at the hot hand in sports. Journal of the Philosophy of Sport, 26(1), 79–87.
Hankin, R. K. S., & Bunker, R. (2016). Bogey teams in sport. Technical report, Auckland University of Technology.
Kovalchik, S. A. (2016). Searching for the GOAT of tennis win prediction. Journal of Quantitative Analysis in Sports, 12(3), 127–138.
Leitner, C., Zeileis, A., & Hornik, K. (2009). Is Federer stronger in a tournament without Nadal? An evaluation of odds and seedings for Wimbledon 2009. Austrian Journal of Statistics, 38(4), 277–286.
Bunker et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 43-64 (2024)
Niall, J. (2013). It’s Murray v Djokovic. The Sydney Morning Herald. https://www.smh.com.au/sport/tennis/its-murray-vdjokovic-20130125-2dcuv.html
Peel, L., & Clauset, A. (2015).Predicting sports scoringdynamics with restoration and anti-persistence. In C. Aggarwal, Z. Zhou, A. Tuzhilin, H. Xiong, & X. Wu (Eds.), 2015 IEEE International Conference on Data Mining (pp. 339–348). IEEE. https://doi.org/10.1109/ICDM36327.2015
Poulton, E. (2004). Mediated patriot games: The construction and representation of national identities in the British television production of Euro’96. International Reviewforthe Sociology of Sport, 39(4), 437–455.
Raab, M. (2012). Simple heuristics in sports. International Review of Sport and Exercise Psychology, 5(2), 104–120.
Steeger, G. M., Dulin, J. L., & Gonzalez, G. O. (2021). Winning and losing streaks in the National Hockey League: Are teams experiencing momentum or are games a sequence of random events? Journal of Quantitative Analysis in Sports, 17(3), 155–170.
Stone, D. F. (2012). Measurement error and the hot hand. The American Statistician, 66(1), 61–66.
Vaughan Williams, L., Liu, C., Dixon, L., & Gerrard, H. (2021). How well do Elo-based ratings predict professional tennis matches? Journal of Quantitative Analysis in Sports, 17(2), 91–105.
Wilms, S. (2014). Mental Training im Sport: Möglichkeiten der Anwendung in der Sozialen Arbeit [Doctoral dissertation, Hochschule für angewandteWissenschaften Hamburg]. HAW Hamburg Repository. https://reposit.hawhamburg.de/handle/20.500.12738/6563
Wood, L. (2017). Jo-Wilfried Tsonga hopes to avoid Australian Open bogey Kei Nishikori. Herald Sun. https://www.heraldsun.com.au/sport/tennis/jowilfriedtsonga-hopes-to-avoid-australian-open-bogey-keinishikori/news-story/2dfaab3f641494447405ae65fc7e7592
Yue, J. C., Chou, E. P., Hsieh, M. H., & Hsiao, L. C. (2022). A study of forecasting tennis matches via the Glicko model. PLoS ONE, 17(4), 1–12 https://doi.org/10.1371/journal.pone.0266838
Bunker et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 43-64 (2024)
Supplementary materials
Supplementary Table 1: Statistically significant (with at least 90% confidence) and bogey player pairs from the ATP tennis dataset, with aggregated Elo ratings-implied probabilities used to compute expected wins.
Player1 (p1) vs player2 (p2) p
G. vs Youzhny M.
M. vs Garcia-Lopez G.
Almagro N. vs Istomin D.
Llodra M. vs Simon G.
Gasquet R. vs Pouille L.
K. vs Berdych T.
Ancic M. vs Blake J.
Acasuso J. vs Rochus C.
Ramirez-Hidalgo R. vs Verdasco F.
D. vs Verdasco F.
Ginepri R. vs Korolev E.
P. vs Mayer L.
Berlocq C. vs Stakhovsky S.
J. vs Gulbis E.
Fognini F. vs Paire B.
Chardy J. vs Isner J.
Querrey S. vs Ramos-Vinolas A.
Isner J. vs Opelka R.
Lorenzi P. vs Paire B.
Notes: As mentioned in the manuscript when describing the proposed method, the FET needs to have a statistically significant p-value but also the expected wins and actual wins need to contradict to be suggestive of the bogey effect between a particular player pair. In Supplementary Table 1 above, there is one case where the expected wins and actual wins did not contradict. In particular, the FET pvalue for Anderson vs Berdych was significant with 1- α = 1 ‒ 0.1 = 90% confidence (p-value = 0.0932 ≤ α), however, the expected wins and actual wins do not contradict. Out of the N = 12 historical matches between Anderson and Berdych in the dataset, Anderson was expected, based on the sums of the respective players’ Elo ratings-implied win probabilities, to win 3.07 of the matches, while Berdych was expected to win 8.93 of the matches. Berdych did better than expected, achieving 12 wins while Anderson achieved zero. Thus, although this was a statistically significant result, since the actual wins and expected wins did not contradict, this is not a case where the bogey effect was deemed to exist.
Supplementary Table 2: Statistically significant (with at least 90% confidence) and bogey player pairs from the ATP tennis dataset, with aggregated betting odds-implied probabilities used to compute expected wins. Player1 (p1) vs player2 (p2)
Djokovic N. vs Monfils G.
R. vs Verdasco F.
J. vs Gulbis E.
N
Y Goffin D. vs Pouille L.
Note: All player pairs in Table 4 are deemed bogey player pairs apart from Djokovic N. vs Monfils G. since the expected and actual wins do not contradict.
Bunker et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 43-64 (2024)
Supplementary Table 3: Statistically significant (with at least 90% confidence) bogey player pairs from the WTA tennis dataset, with aggregated Elo ratings-implied probabilities used to compute expected wins.
Player1 (p1) vs player2 (p2) p
M. vs Williams S.
S. vs Zvonareva V.
A. vs Kuznetsova S.
S. vs Voegele S.
A. vs Radwanska A.
J. vs Stosur S.
Chakvetadze A. vs Jankovic J.
L. vs Shvedova Y.
Hibino N. vs Stosur S.
Krumm K. vs Kirilenko M.
P. vs Radwanska A.
P. vs Johansson M.
A. vs Rybarikova M.
A. vs Wickmayer Y.
D. vs Tsurenko L.
Konta J. vs Strycova B.
Supplementary Table 4: Statistically significant (with at least 90% confidence) bogey player pairs from the WTA tennis dataset, with aggregated Betting odds-implied probabilities used to compute expected wins.
(p1) vs player2 (p2) p
Sharapova M. vs Williams S.
S. vs Zvonareva V.
A. vs Kuznetsova S.
A. vs Radwanska A.
A. vs Jankovic J.
Krumm K. vs Kirilenko M.
P. vs Radwanska A.
Cetkovska P. vs Johansson M.
Riske A. vs Wang Q.
Konta J. vs Strycova B.
Supplementary Table 5: Statistically significant (with at least 90% confidence) bogey player pairs from the ATP tennis Grand Slams dataset, with aggregated Betting odds-implied probabilities used to compute expected wins.
Player1 (p1) vs player2 (p2) p-value FET Expected wins (p1/p2)
Isner J. vs Kohlschreiber P.
Notes: The ATP and WTA Grand Slams datasets with Elo ratings used to compute expected wins, and the WTA Grand Slams dataset with betting odds-implied probabilities used to compute expected wins, all yielded no bogey effect player pairs.
Bunker et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 43-64 (2024)
Supplementary Figure 1: Coefficient of Variation (CV) of Betting Odds and Elo Ratings for each year (A), as well as the more granular view of the mean and CV for each year of Elo rating and Betting Odds (B and C, respectively) for the ATP Grand Slam dataset.
Bunker et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 43-64 (2024)
Supplementary Figure 2: Coefficient of Variation (CV) of Betting Odds and Elo Ratings for each year (A) as well as the more granular view of the mean and CV for each year of Elo rating and Betting Odds (B and C, respectively) for the WTA Grand Slam dataset.
Bunker et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 43-64 (2024)
Supplementary Figure 3: Coefficient of Variation (CV) of Betting Odds and Elo Ratings for each year (A), as well as the more granular view of the mean and CV for each year of Elo rating and Betting Odds (B and C, respectively) for the ATP non-Grand Slam dataset.
Bunker et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 43-64 (2024)
Supplementary Figure 4: Coefficient of Variation (CV) of Betting Odds and Elo Ratings for each year (A), as well as the more granular view of the mean and CV for each year of Elo rating and Betting Odds (B and C, respectively) for the WTA non-Grand Slam dataset.
Supplementary Figure 5: Visualizing, on axes representing the absolute difference in actual wins and expected wins for the player pair, statistically significant (with at least 90% confidence) bogey player pairs from the WTA tennis non-Grand Slams dataset, with aggregated betting odds-implied probabilities used to compute expected wins. The data used to generate this plot is presented below in Supplementary Table 6
Supplementary Table 6: Statistically significant (with at least 90% confidence) bogey player pairs from the WTA tennis non-Grand Slams dataset, with aggregated betting odds-implied probabilities used to compute expected wins. Player1 (p1) vs player2 (p2) p
Stosur S. vs Zvonareva V.
Petkovic A. vs Radwanska A.
Ivanovic A. vs Kuznetsova S. 0.0824 5.29/3.71 9/0
Sharapova M. vs Williams S. 0.0867 3.37/6.63 0/10
Chakvetadze A. vs Jankovic J.
3/0 Y Date Krumm K. vs Kirilenko M.
3/0 Y Ostapenko J. vs Wozniacki C.
Riske A. vs Wang Q.
3/0 Y
Supplementary Figure 6: Visualizing, on axes representing the absolute difference in actual wins and expected wins for the player pair, statistically significant (with at least 90% confidence) bogey player pairs from the WTA non-Grand Slams dataset, with aggregated Elo ratings-implied probabilities used to compute expected wins. The data used to generate this plot is presented below in Supplementary Table 7
Supplementary Table 7: Statistically significant (with at least 90% confidence) bogey player pairs from the WTA non-Grand Slams dataset, with aggregated Elo ratings-implied probabilities used to compute expected wins.
Player1 (p1) vs player2 (p2) p-value
Stosur S. vs Zvonareva V.
Garcia C. vs Ivanovic A.
Stephens S. vs Voegele S.
L. vs Stosur S.
A. vs Radwanska A.
J. vs Stosur S.
A. vs Kuznetsova S.
M. vs Williams S.
A. vs Jankovic J.
N. vs Stosur S.
Krumm K. vs Kirilenko M.
J. vs Wozniacki C.
Petkovic A. vs Puig M.
Supplementary Figure 7: Visualizing, on axes representing the absolute difference in actual wins and expected wins for the player pair, statistically significant (with at least 90% confidence) bogey player pairs from the ATP non-Grand Slams dataset, with aggregated Elo ratings-implied probabilities used to compute expected wins. The data used to generate this plot is presented below in Supplementary Table 8
Supplementary Table 8: Statistically significant (with at least 90% confidence) bogey player pairs from the ATP non-Grand Slams dataset, with aggregated Elo ratings-implied probabilities used to compute expected wins. Player1 (p1) vs player2 (p2) p-value FET
G. vs Youzhny M.
M. vs Garcia-Lopez G.
M. vs Simon G.
N. vs Robredo T.
M. vs Blake J.
J. vs Rochus C.
R. vs Verdasco F.
J. vs Wawrinka S.
R. vs Korolev E.
N. vs Istomin D.
Gulbis E. vs Stepanek R.
P. vs Sijsling I.
J. vs Gulbis E.
Querrey S. vs Ramos-Vinolas A.
Lorenzi P. vs Paire B.
Supplementary Figure 8: Visualizing, on axes representing the absolute difference in actual wins and expected wins for the player pair, statistically significant (with at least 90% confidence) bogey player pairs from the ATP non-Grand Slams dataset, with aggregated betting odds-implied probabilities used to compute expected wins. This figure was generated using the data presented below in Supplementary Table 9
Supplementary Table 9: Statistically significant (with at least 90% confidence) bogey player pairs from the ATP non-Grand Slams dataset, with aggregated betting odds-implied probabilities used to compute expected wins. Player1 (p1) vs player2 (p2) p-value FET
Almagro N. vs Robredo T.
R. vs Verdasco F.
J. vs Wawrinka S.
N. vs Kiefer N.
Kohlschreiber P. vs Sijsling I.
J. vs Gulbis E.
Goffin D. vs Pouille L.
Bunker et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 43-64 (2024)
Supplementary Table 10: Expected win distribution violation quantification for the various datasets. For each type of player pair –whether the player pair contains a bogey player or not – average expected wins and average actual wins were calculated and scaled based on the total number of historical matches between the player pair. Taking the difference between these two values provides a means of quantifying the degree to which the expected win distribution is violated. This data was used to generate the plot depicted in Figure 11 Dataset / method Player pair
/ Elo
ATP / Odds
ATP Grand Slam / Elo
ATP Non-Grand Slam / Elo
w/ bogey player
Grand Slam / Odds
WTA / Elo
WTA / Odds
WTA Grand Slam / Elo
WTA Non-Grand Slam / Elo
WTA Grand Slam / Odds
WTA Non-Grand Slam / Odds
The Journal of Sport and Exercise Science, Vol. 8, Issue 1, 65-73 (2024)
www.jses.net
2703-240X
Youth basketball in New Zealand: Establishing performance norms in the context of 'Balance is Better'
Lee-Anne Taylor1* , Patrick Lander1 , Russell Rayner1,2
1Eastern Institute of Technology – Te Aho Māui, Te Kura Kaupapa Hauora, Hākinakina – School of Health and Sport Science, New Zealand
2Charles Sturt University, School of Allied Health, Exercise and Sports Sciences, Australia
Received: 04.12.2023
Accepted: 04.10.2024
Online: 22.11.2024
Keywords: Youth Adolescent Basketball
Normative data
New Zealand
This study establishes normative data for adolescent basketball players in New Zealand, and seeks to understand if sport participation volumes align with the 'Balance is Better' initiative by Sport New Zealand. The study recruited a convenience sample of 55 junior representative basketball players from Hawke’s Bay, New Zealand, aged 13 to 18 years, comprising 42 males and 13 females. Gender was self-reported in a baseline questionnaire that also covered sports affiliations, weekly training, game hours, and injury history. To determine physical progression through age, participants underwent anthropometric and physical performance assessments, including tests of strength endurance, power, speed, agility, and dynamic balance. Differences between age group data were assessed through one-way ANOVA and non-parametric tests. The findings indicate general adherence to ‘Balance is Better’ recommended activity hours across age groups. Notably, females displayed a decrease in weekly sport hours with age, contrary to male athletes. Female players also engaged more in other sports than males, suggesting less basketball specialisation, especially in the U17 category. Gender differences were evident in physical performance: females showed non-significant changes in strength endurance, jump performance, speed, and agility with age, while males exhibited significant improvements in strength endurance (press-up, p = 0.002; prone hold, p = 0.002; right side hold, p = 0.004; left side hold, p = 0.005), vertical jumps (right, p = 0.011; left, p = 0.019). It is possible that comparisons of female data were unable to detect significant differences owing to low participant numbers. The study reveals that physical performance in youth basketball in New Zealand does not uniformly improve with age. Gender disparities are evident, with females participating in varied sports and males tending towards early basketball specialisation. Overall, participation volumes align with the 'Balance is Better' guidelines.
1. Introduction
Basketball is a dynamic sport requiring strength, endurance, and speed qualities. Prior to the COVID-19 pandemic, basketball was the fastest-growing secondary school sport in New Zealand with a 45% increase in participation over the preceding decade (School
Sport New Zealand, 2018). This growth continued post-pandemic, with basketball inNew Zealand surpassingbothRugby Union and Football (Soccer) in overall participation numbers, second only to Netball (School Sport New Zealand, 2022). Indeed, the 2022 Secondary School Sport New Zealand Census showed a total of 25,387 adolescents played basketball that year, an increase of 62%
*Corresponding Author: Lee-Anne Taylor, School of Health and Sport Science, Eastern Institute of Technology Te Pūkenga, New Zealand, ltaylor@eit.ac.nz
since 2000 when records began (School Sport New Zealand, 2022). This participation growth has been attributed to the global popularity of basketball and increased opportunities for participation across the country, such as tournament play and three-on-three modified games (Basketball New Zealand, 2021a).
Physical performance can vary within individuals as they mature (Baxter-Jones, 2019).Cardiovascular and musculoskeletal capabilities improve with age and maturation, particularly during puberty (Mountjoy et al., 2008). However, these developments are neither linear, nor uniform across individuals. While there are datasets of physical performance and anthropometric data profiles of elite-level basketball players, there is a considerable lack of data on adolescents (Abdelkrim et al., 2010; Benis et al., 2016; Fort-Vanmeerhaeghe et al., 2016; Torres-Unda et al., 2013; Torres-Unda et al., 2016). It is therefore pertinent to collate normative data on adolescent basketball players to assess their physical performance progression as they age, and their potential progressionintoeliteprogrammes.Todate,therearenoagegroup performance targets published by Basketball New Zealand, nor age group development guidelines encompassing both basketballspecific and non-basketball activities such as strength, speed, and cardiovascular development.
In addition to elite performance profiling, normative data can also be used in the management of workloads to moderate the risks of overload and overtraining (Woods et al., 2017). Periods of regional representation are particularly concerning from an overload perspective as the combination of school, community, and elite sports often forces athletes to exceed age-appropriate workload thresholds. Caution has previously been advised in this space for multiple sports (Phibbs et al., 2018; Temm et al., 2022). Thus, investigations into performance profiles of age group athletes are important not only for benchmarking, but also for the prevention of overload injuries.
‘Balance is Better’ is an initiative by Sport New Zealand to assist in creating positive sporting experiences for all New Zealand youth, to keep them active, and in sport. The initiative provides guidance on the amount of training and competition load in alignment with the World Health Organisation (WHO) and the New Zealand Ministry of Health (MoH) guidelines. Furthermore, this initiative brings into question the management of workload across multiple sports and the potential for overloading caused by early sport specialisation. ‘Balance is Better’ suggests that total sport participation hours per week should not exceed the athlete's chronological age irrespective of the number of sports played (Sport New Zealand, 2021). Basketball New Zealand, alongside other major sporting codes, signed an official statement of intent to support the principles of ‘Balance is Better’ in April 2021, with the stated aim of “supporting young people to play multiple sports and raising awareness of the risks of overtraining and overloading” (Basketball New Zealand, 2021b)
Investigations using other sports have shown considerable differences in performance profiles across age group athletes (Taylor & Lander, 2018); however, there is limited profiling of youth basketball players, particularly in New Zealand. Thus, the aim of this study was to report the physical characteristics of adolescent representative basketball players in Hawke’s Bay, New Zealand and compare their workloads with the nationally identified ‘Balance is Better’ philosophy.
2. Methods
2.1. Participants
All players from the Basketball Hawke’s Bay (BBHB) representative age group under-15 (U15), under-17 (U17), and under-19 (U19) years teams from 2020 and 2021 were invited to participate in this study. Fifty-five participants (42 males and 13 females) aged 13 to 18 years agreed to participate from a potential population of sixty-seven participants. The study was approved by the Research Ethics and Approvals Committee of Eastern Institute of Technology, New Zealand. All participants under 16 years of age obtained parental or caregiver consent.
2.2. Data collection
2.2.1.
Baseline questionnaire
On arrival to the testing session, participants completed a baseline questionnaire. This questionnaire collected data on their affiliations with basketball teams, academies, and other sports teams. It also inquired about the athletes’ weekly training and game hours, along with their history of injuries. Weekly training and peak game-time hours were calculated for each participant's engagement in basketball and other sports. Total weekly physical activity hours were derived by summing basketball-related and other sports-related training and game hours.
Potential total basketball hours for secondary school training andgames werederivedbymultiplying theweeklyplayer training and game hours over an 18-week period, corresponding to the scheduled BBHB competition weeks (Basketball Hawke's Bay, 2020a).Assecondaryschoolshavebothjuniorandseniornational tournaments, participants indicated which they were involved in, and maximal game time was calculated on the maximum minutes per game for each of the scheduled tournaments. This prospective approach was also adopted when calculating other sports hours and Physical Education classes.
Basketball Academy training in Hawke’s Bay runs for approximately eight weeks per school term. Players reported how many terms they attended, and total academy time was calculated by multiplying the weekly training duration by the number of weeks attended (Basketball Hawke's Bay, 2020b). If players were also involved in an adult basketball club, club training and game total basketball times were calculated by the player training and game times per week over eight weeks (Basketball Hawke's Bay, 2020c). The total seasonal basketball hours were derived from the sum of all basketball-related activities per participant throughout the season.
Regional representative basketball programmes were calculated from the player training hours over 20 weeks; the number of weeks training for Easter and National tournaments. Game time was calculated on the maximum minutes per game for each of the age group related scheduled tournaments (Easter tournament [U15, U17, U19], June tournament [U19], and July tournaments [U15, U17]; Basketball Hawke's Bay, 2020d)
To create performance profiles, physical testing was conducted in February of 2020 and 2021, following the trialling and selection of all representative teams.
2.2.2. Physical performance testing
Anthropometric measurements, including arm span, leg length, height, and weight were recorded. Body mass index (BMI) was calculated using the formula:
Any player who disclosed a current injury was assessed by a registered physiotherapist and cleared to participate in all performance tests which were not affected by their injury. This screening resulted in no more than two athletes abstaining from various performance tests dependent upon each individual’s injury.
Prior to testing, participants performed a five-minute standardised warm-up, the testing protocol then comprised of 50 minutes of physical performance testing followed by a fiveminute cool-down. A rest period of at least three minutes between each physical performance test was provided for all participants.
Performance testing included a combination of strengthendurance, power, speed, agility and balance tests. The tests used have previously been shown to give an overview of the physical performance characteristics of both adolescent male and female basketball players (Drinkwater et al., 2008; Fort-Vanmeerhaeghe et al., 2016; Wen et al., 2018)
Strength endurance was assessed using prone hold, side holds and maximal press-ups. Prone and side holds required the participant to maintain postural alignment and a stable body position on toes and elbow. Participants were instructed to sustain this alignment for as long as possible; guidance was provided as required. Time to fatigue was recorded in seconds, and the test ended when the participants could no longer maintain postural alignment (Greene et al., 2012; Strand et al., 2014). The maximal press-up test required participants to perform as many full pressups as possible. A valid repetition involved lowering the upper body to a point of 10 cm above the ground which resulted in a minimum elbow flexion of 90 degrees, as per the protocols described by Amasay et al. (2016) and Ryman Augustsson et al. (2009).Thetest ended iftheparticipant failed to;maintain smooth movement without resting at any point, reach the required depth, or maintain a stable position. The total number of press-ups was recorded.
Power testing comprised of a series of vertical jumps (VJ) and broad jumps measured in centimetres. The VJ used a Swift Yardstick 2 vertical jump measurement apparatus (Swift Performance Equipment) with testing procedures identical to those described by Woolford et al. (2013). Participants stood beneath the device and extended their arm directly overhead to determine the starting point. Prior to each VJ, participants were instructed to squat to a comfortable depth and immediately jump as high as possible in one continuous motion while reaching with the designated hand, thereby using a countermovement jump and an arm swing to displace the highest vane of the Yardstick apparatus as described by Castro-Pinero et al. (2009). The performed jumpwasrequiredtobefromtwofeetwithpreliminary steps prohibited. The difference between the starting point and maximal height of each jump was recorded in centimetres. The test was then repeated using the opposite hand. Each jump was
repeated three times with a rest between jumps for a total of six jumps; the best jump from each hand was used for data analysis. For the broad jump, participants completed three standing broad jumps from a fixed starting line. Athletes were permitted a free arm swing and self-selected countermovement jump to maximise their horizontal displacement with procedures identical to those used in Thomas et al. (2020). Participants started with their toes on a start line. The distance was recorded from the starting line to the most posterior portion of the participant's foot closest to the starting line. The furthest jump was used for data analysis.
Sprint testing was conducted over 10 metres (m) using Duo Swift laser timing system, recording 5 m and 10 m splits on the Swift SpeedLight iPad application (Version 493, Swift Performance). Sprint start position was standardised with athletes instructed to begin with their chest as close to the starting beam as possible without touching the beam with the entirety of both feet behind the starting line. The fastest of three trials was recorded for final analysis.
Agility testing used Duo Swift laser timing system, as previously described. Athletes performed a Y-shaped reactive agility test, responding to flashing light stimuli, aiming to complete the test as quickly as possible. Despite the limited transferability of time-based assessments with generic stimuli to sport-specific contexts (Young et al., 2021), time constraints necessitatedtheuseofagenericstimulus. Inthistest,timinglights, activated at 5 m, randomly triggered flashing lights at either the left or right timing gates, necessitating athletes to make an unanticipated 45-degree directional change. Thus, the athletes travelled a total of 10 m with achange of directionat the 5 mmark. To distinguish the capacity for the athlete to make the directional change from sprinting speed, the agility deficit was calculated by subtracting the 0 – 10 m sprint test time from the best agility time, measured in seconds in a procedure previously described by Nimphius et al. (2013).
Dynamic balance was assessed using the Y-dynamic balance test, which encompasses anterior, posterolateral, and posteromedial directional reaches. Composite reach scores for both the right and left sides were derived using a formula previously described by Neves et al. (2017)
2.3. Statistical analysis
All data were assessed using quantile-quantile (QQ) plots for normality and Levene’s test for homogeneity of variance. Residual plots and scatterplots were used to identify any potential outliers. Descriptive statistics (means and standard deviations) were calculated using Microsoft Excel to describe the data collected from each age group. Differences between the groups were determined using the Kruskal-Wallis Test for comparing three groups and Dunn’s test for pairwise comparisons. When comparing between two groups the Mann-Whitney test was used. In addition, theWilcoxon Signed-Rank test was used todetermine if training time in hours exceeded chronological age of participants. All data analyses were performed using R (4.2.3).
To determine the magnitude of differences, Hedge’s g effect sizes were calculated and reported (Bernards et al., 2017; Ellis, 2010) The size of the effect was classified as trivial (< 0.19), small (0.20 – 0.59), moderate (0.60 – 1.19), large (1.20 – 1.99), and very large (2.0 – 4.0; Cohen, 1988; Hopkins et al., 2009)
3. Results
Data were analysed from 55participants, 13 females and 42males. The average age of the female U15 and U17 cohorts was fourteen years, three months, and sixteen years, five months respectively. The anthropometric and physical activity characteristics within the groups of U15 and U17 for females are shown in Table 1 and U15, U17, and U19 for males are shown in Table 2.
All athletes met the training volume guidelines described in the Balance is Better documentation from Sport New Zealand, which recommends that weekly training hours should not exceed an athlete's chronological age in years (Sport New Zealand, 2021). A Wilcoxon signed-rank test confirmed a significant difference between ageandweekly training hours(p <0.001), indicating that this cohort were adhering to these guidelines (Figure 1).
Notes: *Significantly different between
†Moderate effect size ‡Large or Very Large effect size
Table 1: Female anthropometric and physical activity data (M ± SD).
Table 2: Male anthropometric and physical activity data (M ± SD).
Figure 1: Difference between chronological age and weekly training hours for athletes. Notes: Black dot represent data points considered outliers, defined as values more than 1.5 times the interquartile range beyond the quartiles. #Significantly different (p < 0.001)
All participants in the U15 female age group indicated that they had sustained a previous injury at some point in their playing career, 72% of these were either the ankle or foot. The U17 age group had a previous injury rate of 67%, 50% of which were
hand/wrist injuries. Across the two cohorts 42% of U15 females and 17% of U17 females indicated they were currently carrying an injury at the time of testing; however, assessment from a registered physiotherapist identified all female participants were capable of participating in testing.
In the male U15, U17, and U19 cohorts average ages were fourteen years, three months; sixteen years, five months; and seventeen years, seven months, respectively. Previous injuries were identified by 52%, 70%, and 57% of participants in respective age groups. The predominant area the ankle or foot for all age groups, (86%, 86%, and 100%, respectively). Fewer current injuries were indicated by males compared to female athletes, for males these were 8% for U15, 30% for U17, and 14% for U19age groups. Eighty three percent of current injuries within these groups were lower limb injuries. Two players were identified by the physiotherapist as unfit to complete some componentsofthephysicalperformancetesting(dynamicbalance, jumping, press-up, speed, and agility) due to ankle injury.
Comparative data from the female and male U15 and U17 age group physical activities are shown in Table 3. There is a large difference in the number of other sporting teams that female athletes are engaged in compared to males in the U15 (g = 1.13), this difference is greater in the U17 age group (g = 3.21). Despite this, the average weekly total sport hours reduced in females as they aged, by contrast, males' average weekly total sport hours reduced at U17 and then increased in the U19 age group. The female athletes profiled appearto engage inother sports alongside basketball over all age groups while males do not, which may suggest early specialisation of basketball in the U15, U17, and U19 age groups is more prevalent in males than females.
Notes: *Significantly different (p ≤ 0.05). †Moderate effect size. ‡Large or Very large effect size.
Female age group physical performance characteristics are outlined in Table 4, with no significant differences between the age groups. It should be noted that while mean strength characteristics increased with age, there was a reduction in the mean VJ and broad jump with increasing age, along with poorer speed and agility times as athletes moved age groups.
Male age group physical performance characteristics show significant differences between the U15 and U17 groups in all strength endurance tests (prone hold and press-up, both p’s = 0.002) and VJs (right, p = 0.011; left, p = 0.019). Significant differences were also shown between the U15 and U19 in press-
ups (p =0.005)andVJs(right, p =0.003; left, p = 0.001); however, there were no differences between and of the physical performance measures in the U17 and U19 cohorts. Table 5 outlines the male physical performance characteristics. Of potential concern for the female U17 group is the difference in the right and left side anterior direction dynamic balance, with bilateral differences approaching 4 cm (75 cm vs 78.8 cm) since a difference of 4cm or greater is suggested by Neves et al. (2017) to elicit a greater probability of lower limb injury; however, it should be noted these values failed to reach significance (right, p = 0.475; left, p = 0.317).
Table 3: Female vs Male age group comparisons for physical activity data (M ± SD).
4: Female age group physical testing characteristics (M ± SD)
Notes: VJ = vertical jump; R = right; L = left. No significant differences between U15 and U17 groups (p > 0.050). †Moderate effect size. ‡Large or Very large effect size
Notes: VJ = vertical jump; R =right; L =left. *Significant differences between U15 and U17groups (p <0.050). *Significant differences between U17 and U19 groups (p < 0.050). †Moderate effect size. ‡Large or Very large effect size. n-1 Less one participant. n-2 Less two participants.
Table
Table 5: Male age group physical testing characteristics (M ± SD)
4. Discussion
In New Zealand and globally, many sports have focused on understanding athletes' physical development and performance as they mature and advance in their respective sports. Such profiling can provide objective data for talent identification and development (Woods et al., 2017). However, at present, Basketball New Zealand has not published age group performance targets, nor age group development guidelines for basketball-specific and non-basketball activities. To our knowledge, this is the first study in New Zealand to describe the physical characteristics of adolescent representative basketball players in New Zealand and compare their workloads with the national ‘Balance is Better’ philosophy.
One might presume that physical performance increases linearly with age. However, chronological age does not necessarily align with biological maturity nor reflect the accumulation of training and competitive sport experience (Salles et al., 2019). In this study, whilst some of the female physical performance results such as dynamic balance and strength endurance (core and pressup) increased from the U15 to U17 age groups, other physical performance tests such as jump (VJ and broad), speed and agility reduced with age. Notably, except for the prone hold, none of the measures showed significant changes between age groups. For males, improvements were observed in prone hold, side holds from both sides, press-ups, and VJ from U15 to U17. Significant differences between U17 and U19 were seen in prone holds, rightside holds, press-ups, and VJ on both sides. Of note, there was a non-significant reduction in performance of the U19 age group in dynamic balance, speed and agility.
Using data from a similar Netball study in New Zealand this study would suggest that basketball participants are taller and heavier with a greater BMI than netballers (Taylor & Lander, 2018), despite this, when comparing right and left VJ females from the basketball study appeared to jump approximately 15 cm higher than netballers. These differences may reflect sportspecific differences between netball and basketball.
While players, parents, coaches, and sporting organisations should all be aware of performance profile results, it is important to acknowledge that these are a snapshot in time and should not be assumed to increase linearly. Age group comparisons of teams as a collective do not take into consideration individualised biological age markers such as the timing of peak height velocity, and thus we should apply group data to individuals with caution. Additionally, the literature offers limited data on talent identification in basketball (Barraclough et al., 2022) Although derived from a small cohort, this study enhances the understanding of performance profiles in New Zealand basketball.
Research indicates that early sport specialisation may expose children to higher risks of burnout and overuse injuries (Jones & Chang, 2021). Therefore, it is important that youth are provided environments where they can engage in multiple sports, aligning with the 'Balance is Better’ campaign. In contrast, talent identification pathways typically focus on identifying athletes at an early age and nurturing them exclusively within a single sport (Larkin & Reeves, 2018). This approach contradicts the multisportparticipationphilosophyof 'Balance is Better’.The‘Balance is Better’ campaign, developed by Sport New Zealand, advises coaches, parents, and leaders to encourage diverse sports
participation among young people while monitoring participation time. A key guideline suggests aligning weekly participation hours with the child's chronological age (https://balanceisbetter.org.nz/a-practical-guide-for-monitoringathlete-training-and-competition-load/). This study, and others (Standing et al., 2019), would suggest that chronological and biological age are both important maturation metrics to consider for regulating workload.
The data from this research suggests that New Zealand youth basketballers athletes are broadly following the participation ‘Balance is Better' workload guidelines. This is notable given the typical categorisation of basketball as an early specialisation sport (Baker et al., 2019) Despite this, the data suggests a balanced workload among these athletes, aligning with data from High Performance Sport New Zealand, which indicates that athletes generally specialise in their respective sports at 15 years and 5 months (High Performance Sport New Zealand, 2020) Interestingly, the data revealed a gender disparity, showing that male athletes tend to focus exclusively on basketball earlier than female athletes.
Limitations
This study has some limitations that should be considered in future research. First, the dataset used was small and profiled a single region of New Zealand; more research is needed to determine if the findings align with the wider New Zealand youth basketball population. Second, our methods used chronological age as a primary measure; future research should consider indicators of biological age such as the peak height velocity to better understand athlete development on a biological maturation level as well as chronological age (Till et al., 2018)
Conclusion
The aim of this study was to report the physical characteristics of adolescent representative basketball players in Hawke’s Bay, New Zealand, and to compare their workloads with the nationally endorsed ‘Balance is Better’ philosophy. The results are positive, indicating that athletes generally adhere to the recommended participation hours. The study also suggests that male athletes specialise earlier than female athletes, whilst females appear to maintain representation on a range of teams across other sports. However, this observation should be considered in the context of the limited size of the cohort, variable training opportunities, and notable decline in female total sport activity in hours per week. Despitetheselimitationsanditsfocusonasingleregion,thestudy provides a foundation for future research to develop more generalisable guidelines that support the well-being and development of junior basketballers.
Conflict of Interest
The authors declare no conflict of interests.
References
Abdelkrim, N. B., Castagna, C., Jabri, I., Battikh, T., Fazaa, S. E., & Ati, J. E. (2010). Activity profile and physiological
requirements of junior elite basketball players in relation to aerobic-anaerobic fitness. Journal of Strength & Conditioning Research, 24(9), 2330–2342. https://doi.org/JSC.0b013e3181e381c1
Amasay, T., Mier, C., Foley, K., & Carswell, T. (2016). Gender differences in performance of equivalently loaded push-up and bench-press exercises. The Journal of SPORT, 5(1), 46–63. https://doi.org/10.21038/sprt.2016.0513
Baker, J., Wattie, N., & Schorer, J. (2019). A proposed conceptualization of talent in sport: The first step in a long and winding road. Psychology of Sport and Exercise, 43, 27–33. https://doi.org/10.1016/j.psychsport.2018.12.016
Barraclough, S., Till, K., Kerr, A., & Emmonds, S. (2022). Methodological approaches to talent identification in team sports: A narrative review. Sports, 10(6), 1–16. https://doi.org/10.3390/sports10060081
Basketball Hawke's Bay. (2020a). Secondary school competitions. http://www.basketballhawkesbay.co.nz/Competitions1/Secondary
Basketball Hawke's Bay. (2020b). Player development programmes. http://www.basketballhawkesbay.co.nz/PlayerDevelopment-Programmes/Horizons-Basketball-Academy-1 Basketball Hawke's Bay. (2020c). Adult competitions. http://www.basketballhawkesbay.co.nz/Competitions1/Adults-2021
Basketball Hawke's Bay. (2020d). Hawke's Bay reps http://www.basketballhawkesbay.co.nz/Hawkes-Bay-Reps1/U15-U17-Reps
Basketball New Zealand. (2021a). Annual report 2021 Basketball New Zealand. https://nz.basketball/wpcontent/uploads/2022/05/J003979-BBNZ-Annual-Report2021_WEB.pdf
Basketball New Zealand. (2021b). Basketball New Zealand officially signs up to Balance is Better [Press release] https://nz.basketball/basketball-new-zealand-officially-signsup-to-balance-is-better/
Baxter-Jones, A. D. (2019). Physical growth and development in young athletes: Factors of influence and consequence. Kinesiology Review, 8(3), 211–219. https://doi.org/10.1123/kr.2019-0024
Benis, R., Bonato, M., & Torre, A. L. (2016). Elite female basketball players’ body-weight neuromuscular training and performance on the Y-balance test. Journal of Athletic Training, 51(9), 688–695. https://doi.org/10.4085/1062-605051.12.03
Bernards, J. R., Sato, K., Haff, G. G., & Bazyler, C. D. (2017). Current research and statistical practices in sport science and a need for change. Sports (Basel), 5(4). 1–10. https://doi.org/10.3390/sports5040087
Castro-Pinero, J., Gonzalez-Montesinos, J. L., Mora, J., Keating, X. D., Girela-Rejon, M. J., Sjostrom, M., & Ruiz, J. R. (2009). Percentile values for muscular strength field tests in children aged 6 to 17 years: Influence of weight status. Journal of Strength and Conditioning Research, 23(8), 2295–2310. https://doi.org/10.1519/JSC.0b013e3181b8d5c1
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Routledge Academic. Drinkwater, E. J., Pyne, D. B., & McKenna, M. J. (2008). Design and interpretation of anthropometric and fitness testing of
basketball players. Sports Medicine, 38(7), 565–578. https://doi.org/10.2165/00007256-200838070-00004
Ellis, P. D. (2010). The essential guide to effect sizes: Statistical power, meta-analysis, and the interpretation of research results. Cambridge University Press.
Fort-Vanmeerhaeghe, A., Montalvo, A., Latinjak, A., & Unnithan, V. (2016). Physical characteristics of elite adolescent female basketball players and their relationship to match performance. Journal of Human Kinetics, 53, 167–178. https://doi.org/10.1515/hukin-2016-0020
Greene, P. F., Durall, C. J., & Kernozek, T. W. (2012). Intersession reliability and concurrent validity of isometric endurance tests for the lateral trunk muscles. Journal of Sport Rehabilitation, 12(2), 161–166. https://doi.org/10.1123/jsr.21.2.161
High Performance Sport New Zealand. (2020, March 2020). High Performance Sport NZ endorses youth sport campaign. https://hpsnz.org.nz/about-us/news-media/high-performancesport-nz-endorses-youth-sport-campaign/
Hopkins, W. G., Marshall, S. W., Batterham, A. M., & Hanin, J. (2009). Progressive statistics for studies in sports medicine and exercise science. Medicine and Science in Sports and Exercise, 41(1), 3–13. https://doi.org/10.1249/MSS.0b013e31818cb278
Jones, B. R., & Chang, S. H. (2021). The effects of early sport specialization (ESS) on youth sports: Benefits and issues. International Journal of Physical Education, Sports and Health, 8(3), 14–18.
Larkin, P., & Reeves, M. J. (2018). Junior-elite football: Time to re-position talent identification? Soccer & Society, 19(8), 1183–1192. https://doi.org/10.1080/14660970.2018.1432389
Mountjoy, M., Armstrong, N., Bizzini, L., Blimkie, C., Evans, J., Gerrard, D., Hangen, J., Knoll, K., Micheli, L., Sangenis, P., & Van Mechelen, W. (2008). IOC consensus statement: “Training the elite child athlete”. British Journal of Sports Medicine, 42(3), 163–164. https://doi.org/10.1136/bjsm.2007.044016
Neves, L. F., Quadros de Souza, C., Stoffel, M., & Picasso, C. L. M. (2017). The Y Balance Test – How and why to do it? . International Physical Medicine & Rehabilitation Journal, 2(4), 99–100. https://doi.org/10.15406/ipmrj.2017.02.00058
Nimphius,S.,Geib,G.,Spiteri,T.,&Carlisle,D.(2013).“Change of direction” deficit measurement in Division I American football players. Journal of Australian Strength and Conditioning, 21, 115–117.
Phibbs, P. J., Jones, B., Roe, G., Read, D., Darrall-Jones, J., Weakley, J., Rock, A., & Till, K. (2018). The organised chaos of English adolescent rugby union: Influence of weekly match frequency on the variability of match and training loads. European Journal of Sport Science, 18(3), 341–348. https://doi.org/10.1080/17461391.2017.1418026
Ryman Augustsson, S., Bersås, E., Magnusson Thomas, E., Sahlberg, M., Augustsson, J., & Svantesson, U. (2009). Gender differences and reliability of selected physical performance tests in young women and men. Advances in Physiotherapy, 11(2), 64–70. https://doi.org/10.1080/14038190801999679
Salles, W. N., Soares, A. L., Collet, C., Milan, F. J., Palheta, C. E., Mendes, F. G., Kós, L. D., Nascimento, J. V., & Carvalho, H. M. (2019). Influence of age and maturation on perceived
Taylor et al. / The Journal of Sport and Exercise Science, Journal Vol. 8, Issue 1, 65-73 (2024)
collective efficacy in youth basketball players. Cuadernos de Psicología del Deporte, 19(1), 275–282.
School Sport New Zealand. (2018). Basketball the big mover in 2018 School Sport Census https://www.sporty.co.nz/nzsssc/Education/School-SportNZ-Census-Reports
School Sport New Zealand. (2022). School Sport NZ Representation Census 2022 https://www.nzsssc.co.nz/Education/School-Sport-NZCensus-Reports
Sport New Zealand. (2021). Howmuch istoomuch when it comes to youth sport? A guide to understanding specialisation, playing multiple sports, and training load. https://balanceisbetter.org.nz/how-much-is-too-much-whenit-comes-to-youth-sport/.
Standing, R. J., Maulder, P. S., Best, R., & Berger, N. J. (2019). The influence of maturation on functional performance and injury markers in male youth. Cogent Medicine, 6(1), 1–11. https://doi.org/10.1080/2331205X.2019.1632017
Strand, S. L., Hjelm, J., Shoepe, T. C., & Fajardo, M. A. (2014). Norms for an isometric muscle endurance test. Journal of Human Kinetics, 40(1), 93–102. https://doi.org/10.2478/hukin-2014-0011
Swift Performance Equipment. Wacol, QLD, Australia. https://swiftperformance.com/
Taylor, L., & Lander, P. (2018). Adolescent netball players normative data and physical performance profiles. New Zealand Physical Educator, 51(1), 20–24. Temm, D. A., Standing, R. J., & Best, R. (2022). Training, wellbeing and recovery load monitoring in female youth athletes. International Journal of Environmental Research and Public Health, 19(18), 1–21. https://doi.org/10.3390/ijerph191811463
Thomas, E., Petrigna, L., Tabacchi, G., Teixeira, E., Pajaujiene, S., Sturm, D. J., Sahin, F. N., Gómez-López, M., Pausic, J., Paoli, A., Alesi, M., & Bianco, A. (2020). Percentile values of the standing broad jump in children and adolescents aged 6-
18 years old. European Journal of Translation Myology., 30(2), 240–246. https://doi.org/10.4081/ejtm.2019.9050
Till, K., Morris, R., Emmonds, S., Jones, B., & Cobley, S. (2018). Enhancing the evaluation and interpretation of fitness testing data within youth athletes. Strength & Conditioning Journal, 40(5), 24–33. https://doi.org/10.1519/SSC.0000000000000414
Torres-Unda, J., Zarrazquin, I., Gil, J., Ruiz, F., Irazusta, A., Kortajarena, M., Seco, J., & Irazusta, J. (2013). Anthropometric, physiological and maturational characteristics in selected elite and non-elite male adolescent basketball players. Journal of Sports Sciences, 31(2), 196–203. https://doi.org/10.1080/02640414.2012.725133
Torres-Unda, J., Zarrazquin, I., Gravina, L., Zubero, J., Seco, J., Gil, S. M., Gil, J., & Irazusta, J. (2016). Basketball performance is related to maturity and relative age in elite adolescent players. The Journal of Strength & Conditioning Research, 30(5), 1325–1332. https://doi.org/10.1519/jsc.0000000000001224
Wen, N., Dalbo, V. J., Burgos, B., Pyne, D. B., & Scanlan, A. T. (2018). Power testing in basketball: Current practice and future recommendations. The Journal of Strength & Conditioning Research, 32(9), 2677–2691. https://doi.org/10.1519/jsc.0000000000002459
Woods, C. T., Cripps, A., Hopper, L., & Joyce, C. (2017). A comparison of the physical and anthropometric qualities explanatory of talent in the elite junior Australian football development pathway. Journal of Science & Medicine in Sport 20(7), 684–688. https://doi.org/10.1016/j.jsams.2016.11.002
Woolford, S., Polglaze, T., Rowsell, G., & Spencer, M. (2013). Field Testing Principlesand Protocols. In Physiological Tests for Elite Athletes (2nd ed.). Human Kinetics.
Young, W., Rayner, R., & Talpey, S. (2021). It’s time to change direction on agility research: A call to action. Sports Medicine – Open,7(1),1–5.https://doi.org/10.1186/s40798-021-00304y
