Objective Models of sport development often support the assumption that young athletes' psychosocial experiences differ as a result of seemingly minor variations in how their sport activities are designed (eg, participating in team or individual sport; sampling many sports or specialising at an early age). This review was conducted to systematically search sport literature and explore how the design of sport activities relates to psychosocial outcomes.
Design Systematic search, followed by data extraction and synthesis. The Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines were applied and a coding sheet was used to extract article information and code for risk of bias.
Data sources Academic databases and manual search of peer-reviewed journals.
Eligibility criteria for selecting studies Search criteria determined eligibility primarily based on the sample (eg, ages 7 through 17 years) and study design (eg, measured psychosocial constructs).
Results 35 studies were located and were classified within three categories: (1) sport types, (2) sport settings, and (3) individual patterns of sport involvement. These studies represented a wide range of scores when assessed for risk of bias and involved an array of psychosocial constructs, with the most prevalent investigations predicting outcomes such as youth development, self-esteem and depression by comparing (1) team or individual sport participants and (2) youth with varying amounts of sport involvement.
Summary/conclusion As variations in sport activities impact youth sport experiences, it is vital for researchers to carefully describe and study these factors, while practitioners may use the current findings when designing youth sport programmes.
- Sport psychology
Statistics from Altmetric.com
Recognising the need to safeguard the positive experiences afforded through sport participation, advocates have called for numerous changes to the structure of youth sport.1 For example, the Aspen Institute (http://www.aspenprojectplay.org) published a report titled Sport for All Play for Life: A Playbook to Get Every Kid in the Game, which identified eight strategies meant to support early and positive sport experiences for youth. Although many of the strategies involved broad changes to youth sport systems (eg, enhancing coach training), the report specifically recommended four changes to how sport activities are designed, by (1) promoting child-led play, (2) encouraging the sampling of numerous sports, (3) adapting rules, equipment and facilities and (4) considering athlete input when designing activities. Such calls to action represent a growing understanding that simply participating in sport may not be enough to facilitate adaptive developmental outcomes.2 Rather, young athletes may require optimally structured sport activities to have positive sport experiences. With this view in mind, the current review explored existing literature that examined psychosocial outcomes related to participation in differing sport types, settings or patterns of involvement.
Defining sport activities and psychosocial outcomes
‘Youth sport’ often means different things to different people. For example, when queried as to what constitutes youth sport, one person may picture a volunteer-based and grassroots football league organised on community pitches, while another imagines a group of adolescent gymnasts training on a daily basis and selected by professional coaches to attain international success. As a consequence, it is becoming difficult to understand what constitutes ‘ideal’ organised sport activities (ie, those that support healthy youth development). Here, we refer to sport activities as the identifiable or objective actions that comprise an individual's sport involvement, which are shaped by rules, facilities, equipment, normative beliefs and policies. This definition also extends to broader descriptions of sport activities that emerge when researchers borrow ‘ad hoc’ categorisations involving several discrete differences (eg, contrasting competitive and recreational sport), or compare general patterns of activities over a period of time (eg, early sport specialisation).
Broadly speaking, reviews support the notion that the manner in which an activity is designed or structured will invariably influence the outcomes of sport participation. For example, physical maturation,3 injury risk4–6 and expertise development7 ,8 differ across athletes with varying training volumes, sport specialisation pathways and approaches to training. Whereas these outcomes denote influences on performance and health (ie, expertise, injury, maturation), it is equally important to consider how sport activities shape psychosocial experiences. Notably, the International Olympic Committee recently commissioned a consensus statement on healthy development at all levels of youth sport that placed positive psychological experiences and competencies as central components.9 From this perspective, the potential benefits of sport involvement span the many mental processes, self-perceptions and interactions with others that are used to define youth psychosocial well-being (ie, WHO's definition of youth psychosocial well-being10). Although it is clear that sport involvement is associated with numerous psychosocial outcomes in young athletes (ie, ranging from assertiveness and character to perceived well-being11), it is vital to explore the extent to which the design of different sport activities influences psychosocial outcomes, with the hopes of optimising the sport activities that youth engage in.
To better articulate how the outcomes of youth sport are influenced by sport activities, we must consider the differing perspectives researchers use to delineate what an adaptive psychosocial experience entails. As identified by Eime et al,11 many youth sport researchers use positive youth development models.12 These models consider organised sport as a place where young athletes can develop relationships and engage in activities that bolster the development of personal assets such as confidence, competence, character, caring and connectedness. By comparison, numerous other theories also drive youth sport research,13 often exploring achievement goals and self-determination within athletes' motivation. Finally, sport development models (eg, developmental model of sport participation14 ,15) also guide research, by outlining ideal activities and social supports at different stages of an athlete's career. These differing theories and models are the lenses we use to understand youth sport experiences, and have informed applied programmes, interventions and policies that dictate ideal youth sport activities.16–18
The current review
Given the international interest in enhancing youth sport experiences, it is essential to consider the evidence from psychosocial research that informs our understanding of how youth sport activities are designed. However, there are no existing efforts to systematically aggregate empirical findings—often from diverging theoretical backgrounds—regarding how differing types or frequencies of sport activities relate to psychosocial outcomes in youth. Previous reviews have focused on solitary psychosocial issues in relation to activity design (eg, burnout19), and on the psychosocial and physical consequences of specific activity trajectories (eg, specialisation20). However, a comprehensive review is critical for understanding how researchers have studied facets of activity design in relation to psychosocial consequences. As such, the purpose of this review was to systematically search available literature and explore how the design of sport activities relates to psychosocial outcomes. A related objective was to form a conceptual depiction of the ways that youth sport activities are studied to provide a roadmap for researchers and practitioners to study and adapt sport activities to promote ideal psychosocial outcomes.
This systematic review involved original studies published in academic journals, with the goal of identifying research that investigated how different forms of youth sport activities were associated with athletes' psychosocial outcomes. Guided by this objective, this review followed a protocol for conducting reviews that systematically aggregated empirical findings.21 The systematic search protocol was designed a priori and followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines1 and figure 1 illustrates the search in broad terms. The novelty of broadly examining sport activities along with the heterogeneity in constructs nevertheless made it inappropriate to solely pursue an aggregation of evidence. This is particularly the case, as one objective of this review was to explore how sport activities are commonly studied. To adapt to these demands, configurative and scoping review approaches21 ,22 also informed the design of inclusion criteria and the approach to synthesise findings. First, to account for the many ways in which sport activities were operationalised, a relatively generic search was used to be inclusive at the initial stage of the review. Second, applying a comprehensive conceptualisation of psychosocial outcomes was vital because the variety of concepts that had been studied was of equal importance to the nature of the effects. Finally, the data synthesis incorporated approaches to seek aggregation (ie, assessing risk of bias) as well as configuration (ie, classifying sport activities).
The search was conducted in April 2016 in the SPORTDiscus, CINAHL and ERIC via EBSCOhost; MEDLINE via PubMed; PsycINFO via SCOPUS electronic databases. Whereas search strategies were designed uniquely for each database, the search queries combined four groups of terms. Group 1 included the term ‘sport’ and variants. Group 2 included terms related to youth (eg, child, adolescent, teen, boy, girl). Group 3 included psychosocial constructs (eg, psychosocial, social development, youth development, competence, well-being, self-efficacy), with terms derived from relevant theories and recent review articles (eg, systematic review by Eime et al11). Finally, Group 4 included terms commonly used in sport literature to denote types of activities or settings (eg, competitive, recreational, play, practice, competition, individual sport, team sport). The search terms and additional delimiters for each database are provided within online supplementary materials A.
As a supplemental step, we searched Google Scholar using similar search terms to those used with academic databases (ie, first 1000 results were screened at the level of title and abstract) along with a search of the archives of seven peer-reviewed journals (see online supplementary material B for list of journals), and identified studies from the reference lists of relevant reviews and academic articles.
Search results were managed using an EndNote X7 folder, whereby the citation details for all articles (eg, year of publication, authors, journal name, title) were uploaded into a single file. Deduplication (ie, matching and removing duplicate articles) was conducted using online systematic review software (crebp-sra.com).23 The EndNote folder was uploaded to the program to identify duplicates, which were subsequently reviewed by the first author and removed from the database.
Study selection was guided by the inclusion and exclusion criteria listed in table 1. To generate these criteria, it was necessary to establish definitions for the concepts ‘organised sport’, ‘psychosocial constructs’ and ‘sport activities’. Notably, we limited studies to those that primarily investigated organised sports so that broader physical education and recreation environments were not included. Second, this review applied a broad definition of psychological health and well-being from the WHO,10 and as such, included studies that examined concepts related to this definition. As an example of concepts falling outside of this definition, sport-specific expertise (eg, working memory, sport-specific creativity) and behavioural outcomes (eg, dropout) were deemed beyond the scope of the review. Third, the definition of sport activities was based on past research, although we maintained a broad definition to allow ways of studying sport activities to emerge. Finally, it was important to delimit an age range for considering articles that examined youth populations. Although there are diverging opinions of the ages that constitute ‘youth’,25 we sought studies where the sample was primarily 7–17 years of age (ie, 90% of sample was within this range), and where all participants were younger than 20 years of age.
The first and fifth authors (MBE and RB) completed initial screening at the level of article title, followed by titles and abstracts. At this initial level of review, the authors each reviewed half of the entire list of records retrieved. The full texts of the remaining articles were then uploaded, and the inclusion and exclusion criteria were applied independently to all articles by the first and fifth authors, who used an excel file to record decisions. The authors agreed on 88% of the studies (when calculated with Cohen's κ,26 agreement was κ=0.75 (±0.07)). The second author (VA) provided a third review by considering the full text of each study, which afforded an audit of the study selection along with a mechanism to resolve differences in opinion. Finally, included and excluded articles were reviewed by the remainder of the authorship team (KE, LM, JC), who agreed with the consensus attained by the reviewers. Articles excluded at the level of full text are listed within online supplemental materials C.
Data extraction and risk of bias assessment
After identifying eligible studies, the coding tool (see online supplementary file 2) was used by two authors (MBE and VA) as a template for initially recording the information from each article, including: citation details, country of participants, guiding theory, study design, psychosocial constructs studied, sport activity(ies) and related operationalisation, sample characteristics, study findings and analyses. This coding tool was also used to assess risk of bias in relation to study methods, analyses and reporting. An adapted coding tool was developed based on previous reviews.11 ,25 As an example, the Downs and Black27 tool used by Eime et al11 was adapted by removing items that were only relevant for interventions, involving those regarding intervention descriptions, adverse events and blinding participants. In addition to these nine items retained from the Downs and Black tool (ie, items 1, 3, 6, 7, 10, 16, 18, 25, 27), four items regarding how studies operationalised sport activities were added to the methodological items. The resulting coding sheet included 13 items that were each coded using a binary 1/0 rating, primarily spanning methodology (eg, ‘Was the study designed to overcome challenges with correlational designs?’) and analysis (eg, ‘Was sample size adequate for the methods and analytical protocol?’). Greater values on the 13-item scale represented lower risk for bias. Inter-rater reliability for the coding sheet was demonstrated when two authors (MBE and VA) independently coded a subset of six articles with 85% agreement (κ=0.65 (±0.10)). After resolving any disagreements from the initial six articles, the reviewers independently coded a separate group of the remaining articles.
The first step within the synthesis process was to organise studies into groups based on the approach used to study sport activities. The second step was to review studies from each resulting group to summarise evidence. This involved a configurative approach by describing the nature of research and approaches to operationalise activities, along with an aggregative approach to identify the strength of evidence and account for consistency of findings across studies.
Regarding the pool of articles for this review, the database search identified 8169 potentially-relevant citations, and this search was supplemented by identifying 28 studies via manual searches of journals and reference lists, along with an additional 41 studies through a search of Google Scholar. After removing 1300 duplicate records, 6938 articles were screened at the level of title, 292 articles were reviewed at the level of title and abstract and 115 progressed to full-text review. Thirty-five relevant studies were ultimately identified and details of each are included in online supplementary materials D. Study methodologies ranged across cross-sectional designs (n=19), longitudinal designs (including two to four time points; n=12), retrospective methodologies to examine sport histories (n=3) and an observational study of athlete interactions (n=1). Samples ranged from 27 to 13 857 (Median=312) participants who completed surveys in school settings (eg, large school-based studies), or within sport settings as members of sport groups. Research was conducted across North America (ie, USA, Canada), Europe (eg, England, Belgium, Sweden), the South Pacific region (ie, Singapore, Australia) and Africa (ie, Botswana). Assessments of risk of bias in methods and analyses ranged from 7 to 12, with the majority of studies (n=25) fulfilling 9 or more of the risk of bias items.
The initial synthesis process entailed classifying ways of studying activities by grouping together studies that used a similar operationalisation of sport activities, and later placing those subgroups into higher order categories sharing a similar focus. Throughout this process, we classified the research with regard to whether it involved: sport types,28–43 sport settings34 ,35 ,44–49 or patterns of individual involvement.28 ,42 ,43 ,50–62 These three higher order categories are illustrated within figure 2, and each category is described below to synthesise findings. It is important to acknowledge the most predominant psychological theories and models that informed studies in this review—including positive youth development,63–65 deliberate practice,66 achievement goals,67 self-determination,68 moral reasoning69 and the developmental model of sport participation.15 Furthermore, psychosocial constructs included measurement tools representing psychosocial health and well-being (eg, perceived well-being, quality of life, mental health, depression, anxiety, suicidal ideation), sport-specific or general self-concept (eg, self-esteem, self-concept, body dissatisfaction, perceived competence), moral beliefs (eg, moral reasoning, moral decision-making, sportspersonship), development of positive assets (eg, developmental experiences, psychological skills), sport-specific motives and intentions (eg, behavioural regulation, sport enjoyment, intentions to return, psychological need satisfaction, achievement goal orientation) and sport group environments (eg, perceived interdependence, group cohesion).
Articles comparing sport types explored the differences between two or more types of sport, and were generally derived from the rules and conventions that define the competitive and social activities necessary for a given sport. Three sport type contrasts were revealed, relating to (1) amount of interdependence among teammates (n=14), (2) physical contact (ie, contact and non-contact team sports; n=2) and (3) requirements for leanness or aesthetic appeal (n=1).
Foremost within this classification were comparisons based on the amount of interdependence among teammates. Primarily, this was demonstrated by researchers comparing team and individual sport using one of three approaches: (1) self-reported or parent-reported involvement in individual or team sport (ie, categorical placement within one type), (2) self-reported or parent-reported amount of involvement in individual or team sport (ie, number of hours) and (3) recruiting athletes from differing sport types to participate in research. Comparisons between team and individual sport were often made with the expectation that team sport participants would report more positive psychosocial outcomes than those in individual sport, because they are more likely to experience social interactions with teammates and others within sport. Across the published research, there were several studies revealing significant differences, indicating that individual sport athletes reported increased depression38 and social anxiety43 compared with youth involved in team sport. However, other studies reported null differences between the two sport types on these outcomes (eg, depression28), or involved more complex comparisons. Notably, students reporting more frequent team sport involvement demonstrated fewer depressive symptoms,39 and reported higher self-esteem41 than youth who were not involved in sport; however, amount of individual sport involvement was not correlated with these outcomes. Two studies also revealed that youth involved in team and individual sport experienced the most positive psychosocial outcomes (eg, higher quality of life;42 more positive developmental experiences28) or that the effect was moderated by gender.28
The majority of research described above included large school-based samples. In contrast, research conducted within sport settings employed a broad scope of theoretical perspectives to compare individual and team sport—revealing, for example, that team sport athletes had lower ratings of moral reasoning30 ,34 ,35 and perceptions of a mastery orientation.33 Furthermore, a more nuanced operationalisation of degree of interdependence also emerged within sport-based research. Considering the extent that individual sport athletes were required to depend on one another, Donkers and colleagues31 revealed that collectively oriented athletes enjoyed individual sport more, and had stronger intentions to return to sport, when they were required to cooperate during competition (eg, participation in relays).
Whereas the majority of sport type research considered degrees of interdependence, two additional sport type contrasts were also identified: (1) involvement in contact versus low-contact sports (n=2), with one study indicating that contact sport athletes reported increased unsportsperson-like attitudes,40 and (2) participation in sports requiring leanness and/or aesthetic judgements (n=1), where no significant differences were found pertaining to body dissatisfaction for those involved in softball, cross country running and gymnastics.32
The second classification involved comparisons of sport settings, which encompass the differing sport contexts that young athletes participate within. Four sport setting contrasts were revealed, including comparisons between: (1) extracurricular school-based sport and community leagues (n=2), (2) relatively higher or lower levels of competitiveness (n=4), (3) adult-led or peer-led sport (n=1) and (4) experiences in sport leagues that had, or had not, adopted a novel set of rules to support positive experiences (n=1).
First, researchers considering the degree of school or community sport involvement used self-determination and achievement goal theories to develop hypotheses, with studies based in France and Belgium revealing that community sport participants reported a stronger ego-goal orientation compared with school-based participants,47 along with more autonomous motivation for physical activity.44 Level of competitiveness was another central comparison, which involved sport activities that varied with regard to the amount of time commitment, presence of a selection process (eg, try-outs), or broader classification of sport at local, regional or national levels. For example, research focusing on moral development revealed that youth in competitive settings placed greater importance on sporting values such as winning or gamesmanship, compared with those in local or recreational sport.34 ,35
Recent articles expanded the ways that sport settings are studied by: (1) comparing children's experiences while participating in adult-led or peer-led sport practices, with the expectation that peer-led experiences would generate increased learning and enjoyment,45 and (2) comparing youth experiences within rugby leagues based on traditional sport rules with league rules designed to reduce tackling and provide more opportunities to be involved in play.49 Although few significant results in relation to psychosocial outcomes were revealed, these studies faced limitations in sample size or measurement.
Patterns of individual involvement
The final classification represents patterns of athlete involvement, and are framed within the breadth, intensity and timing of an individual's sport participation. Three contrasts involving these patterns were revealed, including assessments of: (1) amount of involvement in organised sport (n=13), (2) amount of sport play and practice throughout a sport career (n=2) and (3) differing specialisation or sampling pathways within sport (n=2).
Amount of sport involvement was the most common approach, driven by expectations that youth who are highly engaged in sport should experience increased psychosocial benefits. Approaches to examining amount of involvement included using self-reported or parent-reported estimations of hours of sport participation per week, or Likert-type rating scales (eg, 1= never; 5= everyday). Studies revealed that students with increased sport involvement reported outcomes such as higher sport competence and self-esteem,60 along with lower depression.52 Nevertheless, links between amount of sport involvement and psychosocial outcomes are likely more complex and bi-directional than indicated in earlier research. For example, by employing path analysis models over a period of 3 years, Adachi and Willoughby50 reported greater support for the predictive relationship of self-esteem to continued sport involvement, than the expected pathway from involvement to self-esteem.
The number of sport teams per year and consecutive years involved in sport were also considered. For example, youth who maintained involvement over two or more years experienced higher quality of life42 and perceived that they made increased contributions to the community61 compared with their less-involved counterparts. To consider the amount of sport involvement and the type of involvement, sports based researchers have studied characteristics of practice measured throughout sport seasons. Using athletes' journals of daily individual deliberate practice (eg, self-directed practice to improve skills), Vink et al59 revealed a positive and reciprocal relationship between deliberate practice and intrinsic motivation. Researchers also examined current athletes' early childhood and adolescent sport involvement through retrospective interviews assessing amounts of sport practice51 or comparing involvement in sport-specific practice and play.54 For example, athletes who accumulated more hours of practice reported higher integrated regulation, although they did not report differences in other forms of motivation, or based on involvement in sport-specific play.54
The final approach to characterising individual involvement is evident within two studies that distinguished athletes according to whether they specialised in a single sport and, if so, the age that specialisation occurred.55 ,58 For example, when compared with recreational athletes or those who specialised at later ages, early specialisers (ie, before 12 years of age) reported higher psychological needs dissatisfaction (although no differences were found with regard to needs satisfaction, mental health or mental illness outcomes).55
Risk of bias
Findings from across these categories should be considered in relation to risk of bias in methods, analyses and reporting. To synthesise this research alongside the variability in risk of bias, table 2 identifies relationships demonstrated by at least two studies in this review, with at least one of those studies fulfilling nine or more risk of bias items; meaning the study had no more than an average risk of bias. Evidence is identified for sport activities related to psychosocial outcomes broadly speaking, along with evidence that relationships exist with specific psychosocial correlates. Regarding positive aspects of study design, studies with the lowest risk of bias applied longitudinal methodologies (eg, daily journaling59) alongside appropriate analyses. These studies also assessed several elements of sport activities (eg, assessing sport type along with amount of involvement), in contrast with large-scale school-based studies that only assessed a single element of sport activity to describe students' sport involvement. A limited number of studies also tested potential mechanisms using mediation models,32 ,37 ,41 that are often vital for identifying what processes were responsible for significant relationships.
Study characteristics impacting risk of bias also varied across the three classifications formed in this review. For example, researchers rarely reported the definition or operationalisation of sport types, and several studies only assigned one sport to each type (eg, basketball compared with soccer29). Regarding individual patterns of involvement, one challenge is that even though constructs like motivation or enjoyment were presumed to result from sport involvement, they may also predict involvement. To account for these types of challenges, studies with lower risk of bias accounted for direction of causality by using repeated measurements or analyses that account for bidirectional influences.59 Finally, there was relatively greater risk of bias among studies comparing sport settings as a result of relying on cross-sectional or retrospective data, vague definitions of sport settings, and designs with limited power.
Variations in youth sport activities can have substantial implications for outcomes such as expertise development, injury risk and physical maturation.70 With the objective of contributing to this understanding, the current review considered how sport activities impact psychosocial outcomes. Most notably, this review generated a conceptual model that consisted of three ways of studying sport activities, including sport types, sport settings and patterns of individual involvement. These conceptual distinctions were, in turn, used as a lens to examine past literature. Perhaps the broadest message obtained through this review is that minor discrepancies in sport activities can substantially impact youth sport experiences. Although factors such as competitiveness or adult involvement could be considered descriptive characteristics, variations in each could shape the psychosocial experiences of young athletes. It is important to note, however, that most effects uncovered in this review were dependent on moderating factors (eg, age or gender), were inconsistent across studies or had only been examined in a limited number of studies. As such, moving forward, careful descriptions of study sport types, settings and patterns of individual involvement are essential to exploring the psychosocial impact of youth sport participation.
Studies examined in this review provide only a partial understanding of how sport activities impact psychosocial outcomes, because of varying definitions of sport activities and diverse outcomes, and because of the complex combination of findings. For instance, several studies support the potential for team sport involvement to positively influence self-esteem and positive development, while mitigating depressive symptoms. However, several studies also revealed that comparisons between individual and team sport are more complex, and may vary across gender or through different combinations of individual or team sport involvement. Furthermore, although expectations of the benefits of team sport often relate to increased social bonds with teammates, this presumption is challenged by two studies that failed to show a correlation between individual or team sport types and team cohesiveness.36 ,38
For researchers hoping to improve the way we study psychosocial outcomes of sport involvement, the variability in findings from past research should be considered in light of the inherent complexity of youth sport activities. Consider, for example, the experiences of a young volleyball player, training 10 hours per week with her school team as well as with a regional competitive club. Each of these characteristics at an individual level may help to explain facets of this athlete's sport activities, yet they should ideally be considered as interacting elements that represent a unique individual experience. Nevertheless, studies primarily considered the impact of solitary variations in sport activities and, at most, the combined influence of two forms of sport activities (eg, one study examined sport involvement over several years separately from team or individual sport participation, but did not analyse combined variation across both variables42). Compared with individual patterns of sport involvement, another challenge is that we often study sport types and settings under the presumption that they represent meaningful differences, although this may not be the case in all situations. As one example, by comparing individual and team sport types solely on the degree of cooperation during competition, researchers overlook potential differences within individual sport—ranging from entirely independent collections of individuals to collective teams that feature collective goals and competition among members.71 In sum, to generate richer representations of youth sport activities, our findings suggest that researchers should establish clearer depictions of sport activities, and use person-centred analytical approaches to account for individual variability when examining their effects (eg, latent class analysis or time series modelling).
Considering limitations in the evidence base, it is also important to consider potential publication bias in existing studies, which is plausibly evident in the over-representation of certain forms of research (eg, ‘positive’ results, research from English-speaking regions). For example, many studies from this review involved large school-based samples that are relatively low-cost, meaning that there is potential to overlook more costly studies that feature richer depictions of sport involvement. In addition, there are methodological limitations of this review. Importantly, it was necessary to adapt existing risk of bias tools to generate a single tool that could be used with varying study designs, and to include novel coding items that reflected the operationalisation of sport activities. This limited our ability to assess the validity of the coding tool, and to benchmark risk of bias in studies from this review against other domains. As such, future reviews should generate a more in-depth reflection on risk of bias or of publication bias.
A broader limitation is the challenge of locating all previous research that has examined sport types, settings and activity patterns. Although our search protocol was designed to locate a spectrum of study designs, the review was challenging for two main reasons. First of all, the variables of interest are descriptive constructs that may not be reported as core study findings. Second, publications spanned many fields of study (eg, sport and developmental psychology, education, sport science, paediatric medicine). Although it is plausible that relevant articles were overlooked, the selection of studies nevertheless provides a representative foundation to support the conceptualisation established in this review.
Despite these limitations, this review provides an opportunity to consider how contemporary research informs the normative behaviours and policies of sport organisations. To do so, it is relevant to reconsider the call to adapt rules, equipment and facilities put forth within the Aspen Institute's ‘Project Play’ introduced earlier in this manuscript. Indeed, sport organisations continue to redesign youth sport leagues,72 and psychosocial models have emerged to guide adaptations to rules and facilities in ways that promote positive sport experiences (eg, competitive engineering17 ,73). However, only one article from this review examined young athletes' experiences following systematic adaptations to their league's sport rules and facilities. Similarly, although psychological theory supports calls to promote child-led sport involvement (ie, free play)74 and sampling of numerous sports,15 there were few cases where these topics were studied in quantitative studies. To align with ongoing shifts in youth sport policy and advocacy, it is vital to build on these findings by exploring psychosocial implications of sport specialisation, adaptations to youth sport rules and the extent that youth are involved in peer-led and adult-led practice and play.75 An improved evidence base may inform decision-making when shaping sport programmes to ensure personal development, as opposed to simply ensuring optimal development of sport-specific skills.18 Enhancing the evidence base is particularly vital because efforts to improve sport activities (eg, changing rules or policies49) are low-cost and can be easily maintained, compared with many other approaches to impact youth sport (eg, coach development76).
For sport medicine practitioners, it is vital to develop awareness of how young athletes' continually changing sport activities may influence indicators of positive psychosocial development and well-being (eg, self-esteem, enjoyment, motivation, moral development, quality of life), as well as indicators of psychological distress (eg, depression, hopelessness, externalising behaviours, social anxiety). Consequently, practitioners should closely monitor how changes in sport involvement (ie, frequency or intensity of involvement; changes in competitive level) impact psychosocial outcomes. For example, similar to the importance of assessing training load placed on athletes,77 ongoing and formalised assessments could monitor clients' psychosocial outcomes such as well-being or intrinsic motivation as they experience changes in the amount of sport involvement or the types of sport activities in which youth participate. Furthermore, the findings of this review could be employed to advocate for change in sport organisations in ways that promote psychosocial well-being. In sum, sport practitioners (along with sport researchers and leaders of sport organisations) can act as central figures in the effort to find ways to structure sport programmes that balance the potential influences of three broad elements that help define youth's sport involvement—sport types, sport settings and patterns of individual involvement.
What is already known?
When youth sport involves optimal environments, it can be a powerful organised activity for supporting healthy development, and can promote an array of psychosocial outcomes.
Outcomes such as expertise development, injury risk and physical maturation are influenced by the nature of the organised sport activities for which youth are involved (ie, identifiable or objective actions that comprise an individual's involvement in sport, such as training volumes, sport specialisation, design of sport practices).
It is vital to examine the ways that sport activities are studied in relation to psychosocial outcomes—to locate evidence that may guide efforts to design youth sport, and to improve this field of research in the future.
What are the new findings?
Thirty-five articles investigated how psychosocial constructs were predicted by differing patterns of youth sport activities. Sport activities varied across sport types (eg, level of interdependence with teammates), sport settings (eg, level of competitiveness) or across individual patterns of involvement.
Considering that effects were often dependent on moderating factors (eg, age or gender), were inconsistent across studies or had only been examined in a limited number of studies, researchers must conduct more detailed assessments and analyses regarding the influence of varying sport types, settings, and activity patterns. Despite these limitations, the body of evidence indicates that psychosocial outcomes of sport involvement may be optimised when youth engage in certain forms of sport activities.
Youth involved in sport groups featuring greater interdependence reported outcomes such as enhanced developmental experiences and self-esteem, as well as lower depression and poorer moral reasoning. These effects were not consistent across all studies, however, and varied across gender and age.
Youth who were more deeply involved in sport overall (ie, increased number of sport teams or higher weekly involvement) reported outcomes such as lower depression and higher self-esteem, although these effects varied across studies and there was evidence that those with very high involvement (eg, >17 hours of sport involvement per week) experienced higher depression.
Twitter Follow Michael Blair Evans @mblairevans
Contributors All authors contributed significantly to the manuscript. The initial idea for the review emerged through discussions of MBE and JC. All authors contributed to the study protocol and design through group discussions. MBE, VA and RB conducted the majority of the article extraction and coding. All authors contributed significantly throughout the writing and preparation process and approved of the submitted version of the manuscript.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.