A Longitudinal Study of Sport Participation and Perceived Social Competence in Youth

Journal of Adolescent Health xxx (2019) 1e8

www.jahonline.org Original article

A Longitudinal Study of Sport Participation and Perceived Social Competence in Youth Chloe Bedard, B.H.Sc. a, b, Steven Hanna, Ph.D. b, c, and John Cairney, M.A., Ph.D. a, d, * a

INfant and Child Health (INCH) Lab, Department of Family Medicine, McMaster University, Hamilton, Ontario, Canada Department of Health Research Methods, Evidence, and Impact, McMaster University, Hamilton, Ontario, Canada School of Rehabilitation Science, McMaster University, Hamilton, Ontario, Canada d INfant and Child Health (INCH) Lab, Faculty of Kinesiology and Physical Education, University of Toronto, Toronto, Ontario, Canada b c

Article history: Received December 10, 2018; Accepted September 20, 2019 Keywords: Physical activity; Sport; Adolescence; Psychosocial health

A B S T R A C T

Purpose: Participation in sport presents a unique setting for positive youth psychosocial development. Evidence supporting the relationship between sport and perceived social competence, however, is lacking longitudinal evidence; therefore, it is unknown how the relationship changes through late childhood to early adolescence. The main objective of this study was to evaluate the association between sport participation and self-perceived social competence over 4 years of early adolescence. Methods: The data for this study are from the longitudinal cohort Physical Health and Activity Study. A total of 2,278 children in grade 4 were followed for 4 years until age 13e14 years. Selfperceived social competence (SPSC) was measured using Harter’s Self-Perception Profile for Children. Sport participation was measured using the Participation Questionnaire. Mixed effects models were used to evaluate the effect of sport as a continuous and categorical variable on SPSC. Results: There was a significant association of sport at baseline (b ¼ .06, 95% CI: .04e.08) and a significant association of sport over time (b ¼ .01, 95% CI: 4  103 to .017) on SPSC for both males and females. When examining sport participation categorically, compared with no sport participation, participation in any category of sport (in-school, out-of-school, or both) is positively associated with SPSC. Conclusions: The results of this study demonstrated that higher participation in sport is associated with small gains in perceptions of social competence during late childhood to early adolescence, suggesting that sport may be a small yet important contributor to young adolescents’ perceptions of their social capabilities. Ó 2019 Society for Adolescent Health and Medicine. All rights reserved.

Conflicts of interest: The authors have no conflicts of interest to declare. The study sponsor did not have any role in the study design, data collection or analysis, writing of the article, or decision to submit the article. C.B. wrote the first draft of the article and did not receive any form of payment. * Address correspondence to: John Cairney, M.A., Ph.D., INfant and Child Health (INCH) Lab, Faculty of Kinesiology and Physical Education, University of Toronto, 100 Main St. West, Toronto, Ontario L8P 1H6, Canada. E-mail address: [email protected] (J. Cairney). 1054-139X/Ó 2019 Society for Adolescent Health and Medicine. All rights reserved. https://doi.org/10.1016/j.jadohealth.2019.09.017

IMPLICATIONS AND CONTRIBUTION

The results of this study show that sport participation is related to feelings of social competence, and this relationship increases across late childhood into early adolescence. Demonstrating this relationship during this developmental period provides evidence that sport is an important contributor to psychosocial development in youth.

Sport presents a unique setting for positive youth psychosocial development because it has been classified as an environment in which youth are more motivated to devote their time, energy, and focus compared with other environments such as school [1]. There are several opportunities in sport for children and youth to develop personal and interpersonal skills that can enhance their well-being. Athletes need to coordinate their

2

C. Bedard et al. / Journal of Adolescent Health xxx (2019) 1e8

movements with their teammates, learn from and communicate with their coaches, resolve conflicts, learn to be accountable to others, and work cohesively as a team to successfully achieve a common goal, all of which are thought to enhance their psychosocial well-being [2]. Limited research has suggested that sport participation is associated with specific psychosocial outcomes such as interpersonal relationships, self-esteem, and overall quality of life, particularly among adolescents [3e7]. However, there are many gaps in the field that warrant further studies. Specifically, the extant literature is largely limited in its ability to quantify change in the effect of sport across time, demonstrate associations with critical indicators of psychosocial well-being, evaluate differential effects of sport context, and control for sociodemographic selection factors. The quantitative evidence in support of the relationship between sport and psychosocial outcomes largely comes from cross-sectional studies [6e10]. Cross-sectional studies, however, cannot evaluate changes over time or estimate the extent of individual differences in change. These limitations are important in the context of sport and psychosocial well-being because the sequence of this relationship remains unclear, and we may expect that the relationship may change over time as children progress through adolescence. Adolescence is typically characterized as a step-wise progression through puberty with accompanied biological, psychological, cognitive, and social changes; maturing girls and boys progress through this transitional period at varying ages and with varying speed [11]. The saliency of sport participation on one’s self-concept and social identity may change with time. As a result, longitudinal studies beginning earlier in adolescence will be critical in specifying the nature of the relationship between sport participation and psychosocial outcomes. The few longitudinal studies that do exist in this field report somewhat conflicting results and have methodological limitations that reduce their utility in the study of this relationship over time. Dimech and Seiler [12] and Findlay and Coplan [13] longitudinally examined the relationship between sport participation and social anxiety and social skills, respectively, in primary schooleaged children. Although Dimech and Seiler found a significant effect of team sport participation reducing levels of social anxiety over a 1-year period, Findlay and Coplan [13] did not detect any relationship over 1 year between sport and social skills or positive adjustment. The discrepant findings may be due in part to different [14] measurements of both sport participation and the type of outcome. In addition, both studies sampled fewer than 250 participants and only twice, 1 year apart. This short period may not be sufficiently long to examine changes, particularly over adolescence, which often takes place over durations exceeding 1 year [11]. Marsh [15] and Viau et al. [16] conducted cohort studies over longer periods of adolescence examining extracurricular activities, including sport and the relationship with psychological variables. Marsh [15] studied older adolescents in their final 2 years of high school and 2 years postgraduation and demonstrated that greater participation in various extracurricular activities was associated with positive self-concept and found that sport, in particular, had a large influence on social self-concept. Similarly, Viau et al. [16] tracked extracurricular participation rates over high school in adolescents participating in at least one extracurricular activity to examine effect on adjustment variables measured at age 18 years; the authors found a positive relationship between duration of participation and later adjustment. Although both

longitudinal, neither study is able to describe changes in the relationship throughout late childhood to early adolescence (roughly defined as 9e14 years), a dynamic period of time preceding typical sport dropout [17]. Examining changes across time among young adolescents will be valuable to understand the impact of sport at a time before children begin to withdraw from participation. Another limitation of the extant literature is the paucity of research examining a critical indicator of psychosocial wellbeing, perceived social competence. Perceived social competence is defined as one’s evaluation of their own abilities to engage successfully in social interactions with others [18]. Perceived social competence is positively related to resilience [19], quality of interactions and relationships with others [20], the ability to seek and sustain social support [21], and is negatively related to mental illness [22]. Therefore, evaluation of the impact of sport on ratings of perceived social competence may then elucidate the role of sport as a possible target intervention to improve general mental health and promote prosocial skill development. In addition, it is important to consider the role of selection factors such as sex and socioeconomic status in the relationship between sport and perceived social competence. There appears mixed evidence on the direction of sex effects on rating of social self-concept [15,23]; however, girls consistently report lower participation in physical activity and sports in particular [24]. Considering the potential promotive influence of sport on the development of positive self-concept, it is critical to characterize any differential effect by sex. Given the financial requirement of sport participation, particularly with out-of-school sport (OSS), it is also critical to evaluate the effect of socioeconomic status within the relationship between sport and psychosocial well-being. Children from socioeconomically advantaged families may have more opportunities to engage in sport activities and have other advantages that foster positive self-concept, therefore representing a potential confounding variable. Finally, the context of the sport activity, whether it takes place within the school or outside of school, may be relevant to the relationship between participation and perceived social competence [14]. School-based sports have different attributes than out of school sports, which may exert different effects on selfconcept. For example, school sports are traditionally encouraged and promoted within school as a socially sanctioned activity; therefore, according to social control theory, participation on the team would increase feelings of attachment to the school, which, in turn, increases feelings of self-worth and positive self-concept [25]. Furthermore, school-based sports are often available to students at no (or low) costs; therefore, socioeconomic status may be less of a barrier to participation. Conversely, participation in out of school sports can expose children to different peer groups and provide opportunities to challenge and develop their social skills. However, OSS often presents a financial barrier to socioeconomically disadvantaged families. Guèvremont et al. [14] found that participation in only in-school sports (ISSs) had different associations to socioemotional health compared with participation in both in-school and out-of-school (BIOS) and only OSS in high school students, suggesting that the out of school sport context may be conferring particular benefits to mental well-being. However, it is still unknown how these particular sport contexts affect younger adolescents over time.

C. Bedard et al. / Journal of Adolescent Health xxx (2019) 1e8

3

Table 1 Participant demographics over time

Sample size Sexa Male Female Ageb Household incomeb ($) SPSCb Sportb

Wave 2

Wave 3

Wave 4

Wave 5

Wave 6

Wave 7

Wave 8

Spring 2005

Fall 2005

Spring 2006

Fall 2006

Spring 2007

Fall 2007

Fall 2008

2,278

2,228

2,273

2,134

2,141

1,898

1,805

962 934 12.4 d 19.6 4.3

913 892 13.4 d 19.3 4.0

1,158 1,120 9.9 61,307 18.5 5.6

(50.2) (49.2) (.4) (19,067) (4.3) (4.9)

1,130 1,198 10.3 d 18.6 4.6

(50.7) (49.3) (.4) (4.1) (4.6)

1,159 1,114 10.9 d 19.2 5.8

(51.0) (49.0) (.3) (4.1) (5.2)

1,076 1,058 11.3 d 19.4 4.3

(50.4) (49.6) (.3) (4.1) (4.3)

1,079 1,062 11.9 d 19.6 5.6

(50.4) (49.6) (.3) (4.0) (5.1)

(50.7) (49.3) (.3) (3.8) (4.3)

(50.6) (49.4) (.3) (3.7) (4.0)

SPSC ¼ self-perceived social competence. a Data are presented as n (%). b Data are presented as means (standard deviations).

The present study was designed to fill the aforementioned knowledge gaps in the literature and overcome the methodological limitations and contradictory results of current research. Therefore, the main objectives of this proposed study are as follows: (1) to evaluate the association between sport participation and self-perceived social competence (SPSC) in children longitudinally over 4 years between late childhood and early adolescence; (2) to evaluate the effect of sex and socioeconomic status on the relationship between sport participation and perceived social competence longitudinally between late childhood and early adolescence; and (3) to evaluate the differential impact of ISS and OSS activities on perceived social competence longitudinally between late childhood and early adolescence. Methods Participants The data for this study are from the Physical Health and Activity Study Team (PHAST) project. Further description of the PHAST design and procedures have been described elsewhere [26]. Briefly, PHAST is a longitudinal cohort study conducted in 2005e2009 that aimed to examine PA participation, health, and psychosocial trends of a typically developing population of children. The study recruited children in grade 4, aged 9e 10 years, who were enrolled in the public school system in the Niagara region of Ontario, Canada, and followed them over 4 years to age 13e14 years. Eligibility criteria required children to be able to read and speak English and not to be diagnosed with any physical or intellectual disabilities. PHAST invited all 92 schools in the district to be involved in the study, of which 75 agreed to participate. Within these 75 schools, a large sample of 2,278 children were recruited (95.8% of all children in the schools). Of these 2,278 children, approximately 69% provided data for the full follow-up period for a total of 1,581 participants across eight waves of data collection. All enrolled children provided written and informed assent, and written and informed consent was obtained from parents of participating children. PHAST received ethical approval from Brock University and the District School Board of Niagara. Study procedures A pilot phase of data collection took place in the Fall of 2004 (wave 1). Subsequent data collection took place biannually for the first 3 years (Spring and Fall of 2005e2007) and annually

thereafter (Fall of 2008 and 2009). Trained research assistants collected data from participating children during regular school hours on demographic variables, sport participation, and selfperceived competence. Because a different set of questionnaires was administered once the participants entered into high school during the final wave of data collection (wave 9), only waves 2e8 were included in the analysis. Descriptive statistics of key measures are provided in Table 1. Measures Self-perceived social competence The dependent variable, SPSC, was measured using Harter’s Self-Perception Profile for Children [27]. This self-report survey asks children to evaluate their competence in several domains, including academics, social competence, athletics, physical appearance, behavior, and overall global self-worth. Scores range from 1 to 4 for each item, and each subscale consists of six items. There is evidence of good internal consistency (Cronbach’s alphas for the SPSC subscale ranging from .75 to .84 among children aged 8e14 years in the U.S.) and construct validity (significant negative correlations between the SPSC subscale and social phobias, r ¼ .55, and trait anxiety, r ¼ .51, among children aged 8e14 years from the Netherlands) [28,29]. This questionnaire was administered to participants at waves 2 to 8. The scores on the items pertaining to SPSC subscale were summed to create the dependent variable ranging from 6 to 24, with higher values representing higher perceived social competence. Participation in sport The primary independent variable is sport participation as measured using the Participation Questionnaire (PQ) [30]. There are six items that assess participation in organized sport through intramurals, ISS teams, OSS teams, and sport clubs and lessons in the current school year and over the last 12 months (e.g., “this school year, how many school sports teams have you belonged to?”). The six items are summed with total scores ranging from 0 to 29; a score of 0 represents no sport (NS) participation, and higher scores represent participation in a greater number of sports (e.g., multiple sport teams and/or clubs). These six items may also be scored categorically as ISS and OSS. The PQ has evidence of testeretest reliability over 2 weeks (r ¼ .81) and construct validity as evidenced by significant association with

4

C. Bedard et al. / Journal of Adolescent Health xxx (2019) 1e8

teacher’s assessment of PA (r ¼ .62) in a sample of children in grades 4 to 8 in Ontario, Canada [30].

Children were administered a basic demographic questionnaire at baseline in which they were asked to report their date of birth, sex, and the last four digits of their postal code. To obtain an estimate of the household income for each child, the last four digits of the reported postal code was matched to the 2006 Canadian census data to obtain an estimate of the median household income for specific geographical dissemination areas [31].

with the significant covariates from models 3 and 4. Finally, model 6 included interaction terms between time and each category of sport participation. Models 5 and 6 did not include a random slope for school because the absolute correlation between the random slope and intercept was greater than .95; therefore, a more parsimonious model was specified. Models were compared using a general likelihood ratio test testing the change in deviance (2Log Likelihood) using a chi-square distribution with fit statistics estimated using the full maximum likelihood method. All tests of statistical significance are two tailed and will be assessed against a criterion of p < .05. All statistical tests were conducted using SAS version 9.4 SAS Institute (Canada) Inc.

Missing data

Results

PHAST was a school-based study, and as such, the attrition over the study period was mostly due to children being absent from school on the day of data collection. There were three main patterns of missing data: complete data (i.e., data collected at all eight waves; n ¼ 1,294, 56.8%), missing the last two waves of data collection (n ¼ 565, 24.8%), missing the last 4e6 waves (n ¼ 257, 11.3%), and missing randomly across the study (n ¼ 162, 7.1%). A Spearman’s rho correlation between baseline SPSC scores (the dependent variable) and the number of waves completed revealed an association of .053; although significant, this correlation is very weak. Similarly, a Spearman’s rho correlation between baseline sport participation and the number of completed waves was .002 (p > .05). T-tests and chi-square tests assessing differences in each covariate comparing those with complete and incomplete SPSC data showed some significant differences at various waves; however, the pattern of differences was inconsistent. For example, at wave 2, those with complete SPSC data participated in more sport activities; however at wave 6, those with complete SPSC data participated in fewer sport activities; similarly, at wave 4, those with complete SPSC data were younger, but at wave 7, those with completed SPSC tended to be older. Therefore, we assume that these missing data did not systematically bias the results. Missing data were not imputed.

Table 1 displays the descriptive statistics of the participants over each wave of data collection. Model 1 (Table 2) shows there is a significant association between time and SPSC such that approximately every 6 months (wave 1), there was an increase of .18 points on the SPSC subscale (p < .0001). Results of model 1 also show that there is significant variability in both the intercept and slope across schools and students. There is a significant negative correlation between the slopes and intercepts, indicating that the effect of time on SPSC is greater for schools with a lower average level of SPSC and students with lower baseline SPSC. Finally, model 1 shows that there is significant variability within participants not explained by time. In model 2 (Table 2), sport participation has a small significant association with SPSC over time (p < .0001). Model 3 (Table 2) shows that household income has a small independent association with SPSC (p ¼ .02), with the association with sport participation unchanged. The change in deviance between each of these models was statistically significant (p < .001). The fixed effect of sex and interactions with sex are not statistically significant. Figure 1 shows the similar effect of sport participation over time for males and females. Model 5 (Table 3) shows that compared with NS, participation in any category of sport (ISS, OSS, and BIOS) is significantly associated with SPSC. The change in deviance between models 5 and 6, however, is not statistically significant, suggesting that interaction between time and type of sport participation is not significant overall.

Demographic characteristics

Statistical analyses Mixed effects models were used to evaluate change in SPSC within children. Random intercepts and slopes were included for both school and student. The covariance parameters were estimated using the restricted maximum likelihood estimation, and an unstructured covariance matrix was specified. The Satterthwaite equation was used to compute the degrees of freedom for the significance test of the fixed effects parameters. Time was included in each model as “waves” and was centered so that the intercept would correspond to baseline data at wave 2. Model 1 regressed the SPSC variable onto time, providing a basic developmental model of SPSC over time across late childhood and early adolescence. In model 2, the fixed effect of sport participation, as a continuous variable, and the interaction term between sport and time were added to the model. Model 3 added household income as a covariate. Model 4 tested the moderating effect of sex by including sex by sport, sex by time, and sex by sport by time interaction terms. A fifth model was built with sport participation as a categorical variable divided into NS participation, ISS participation only, OSS participation only, and BIOS participation

Discussion Summary of principle findings The objective of the present study was to examine the association of sport participation on perceived social competence in males and females over 4 years during the transition from late childhood to early adolescence. The results of the developmental model show that in grade 4, students vary widely in their perceptions of their social competence, and this corresponds with varying rates of change over 4 years. Unsurprisingly, those students with lower levels of SPSC showed greater improvements over time. Models 2e4 show that sport participation has a small positive association with both male’s and female’s perceptions of their social competence, and this relationship increases slightly over time. The influence of sport on SPSC is largest when children participate in BIOS compared with NS participation. However, participation in any type of

C. Bedard et al. / Journal of Adolescent Health xxx (2019) 1e8

5

Table 2 Mixed effect model of the trend in perceived social competence over time by sport participation, household income, and sex

Intercept Time (waves)

Model 1 Estimate (SE)

Model 2 Estimate (SE)

Model 3 Estimate (SE)

Model 4 Estimate (SE)

18.6 (.10)** 95% CI: 18.38 to 18.80 .18 (.02)** 95% CI: .14 to .22

18.3 (.12)** 95% CI: 18.0 to 18.5 .14 (.02)** 95% CI: .09 to .19 .06 (.01)** 95% CI: .03 to .08 1.1 (3.0  103)** 95% CI: 6  103 to .02

17.9 (.23)** 95% CI: 17.5 to 18.4 .14 (.03)** 95% CI: .09 to .20 .06 (.01)** 95% CI: .04 to .08 .01 (3.2  103)* 95% CI: 4  103 to .017 7.0  103 (3.2  103)* 95% CI: 6  104 to .01

18 .1 (.25)** 95% CI:17.6 to 18.6 .12 (.04)** 95% CI: .05 to .19 .05 (.01)* 95% CI: .01 to .07 1.1 (4.5  103)* 95% CI: 2  103 to .02 7.0  103 (3.3  103)* 95% CI: 6  104 to .01 .22 (.20) 95% CI: .6 to .18 .05 (.04) 95% CI: .04 to .13 1.2 (.02) 95% CI: .02 to .07 2.4  103 (6.3  103) 95% CI: .01 to 9  103 .51* .12* .89*

Sport Sport *time Income Sex Sex  time Sex  sport Sex  sport  time SD of school-intercept SD of school-slope Correlation between random effects of the school SD of student-intercept SD of student-slope Correlation between random effects of the students SD of the residual 2LLa

.55* .10* .83*

.54* .10* .84*

.51* .12* .90*

3.58** .53** .51**

3.52** .52** .53**

3.46** .52** .52**

3.46** .52** .52**

2.34** 71401.8

2.34** 65613.5

2.34** 65611.1

2.34 71588.8

CI ¼ confidence interval; SE ¼ standard error. *p < .05; **p < .0001. a Computed from a model using full maximum likelihood method.

sport was also significantly associated with SPSC; this relationship does not appear to change over time. The standard deviation within students remained large and fairly stable across each of the models demonstrating that time and sport do not fully explain variability of SPSC within students.

more relevant to SPSC rather than the number of sport activities in which the student takes part. Nonetheless, given both the small size of the effect and observational design of the study, the results should be interpreted with caution. Comparison with extant literature

Practical significance One of the limitations of the SPSC subscale is the lack of a defined minimal important difference threshold, thereby making practical interpretations difficult. However, practical significance may be indirectly estimated by examining differences between populations of young adolescents with typical development and clinical populations with known deficiencies in social selfconcept. For example, cross-sectional comparisons between children with typical development and children with motor coordination difficulties show significant mean differences on the SPSC scale between 1.38 and 2.88 at age 6e10 years [32]. Similarly, older children, aged 12e14 years, with motor coordination difficulties report scores, on average, 1.2 points lower on the SPSC scale, compared with their typically developing peers [32]. From this perspective, the results of the present study show that the effect of participating in three sport activities (the median number of sports) each year will increase their SPSC by a total of 1.02 over 4 yearsdthis could potentially close/narrow the gap between a child with movement challenges and their typically developing peer. In addition, when operationalizing sport with respect to school, the effect of participating in BIOS, compared with NS participation, appears to be larger; this suggests that the specific context in which sport takes place may be

The results of this study are largely consistent with literature examining the development of self-concept over time and its relationship with sport participation. When considering the influence of sport on multiple dimensions of self-concept, it is often suggested that particular domains of self-concept related to physical or athletic ability are more likely to be affected. In support of this view, Zaharopoulos and Hodge [7] found only perceptions of physical competence, not global self-concept, to be associated with participation in athletics among a crosssectional sample of students aged 13e17 years. However, Donaldson and Ronan [6] found sport participation to be positively related to multiple domains of self-concept, including perceptions of social competence in addition to perceptions of athletic competence and global self-worth in a sample of young adolescents aged 11e13 years. The differences between these studies may be partly due to methodological variability with respect to the measurement of self-concept, specificity of sport within the physical activity construct, and the ages of the sample. Donaldson and Ronan recruited a younger sample, more closely resembling the sample used in the present study and found an association between sport participation and SPSC. These findings suggest that sport participation may be important to multiple domains of self-concept during early adolescence compared with

6

C. Bedard et al. / Journal of Adolescent Health xxx (2019) 1e8

Figure 1. Effect of sport on perceived social competence over time by sex.

later adolescence during which time sport may influence more strongly physical or athletic perceptions of self. When considering the impact of any extracurricular activities or clubs such as debate/drama, photography, academic clubs (e.g., science club), student government, and so forth, in addition to sport, the effect on social self-concept is similar, although often larger effects are observed [33]. Linver et al. [10] found that participation in sport and other organized activities was

associated with the highest levels of positive youth development (i.e., academic and social competence, confidence, and prosocial behaviors) compared with no participation in any organized activities. Viau et al. [16] found similar results when examining the impact of participation in organized activities (i.e., sport, cultural, and prosocial activities) on various indicators of adolescent adjustment at age 18 years, such as symptoms of depression, problematic alcohol consumption, and civic

C. Bedard et al. / Journal of Adolescent Health xxx (2019) 1e8 Table 3 Mixed effect model of the trend in perceived social competence over time by sport as categorized by school-involvement

Intercept Time (waves) Income NSa ISS OSS BIOS

Model 5 Estimate (SE)

Model 6 Estimate (SE)

16.90 (.23)** 95% CI: 16.5 to 17.5 .17 (.02)** 95% CI: .13 to .20 4.4  103 (3.3  103) 95% CI: 3  103 to .01 0 1.42 (.30)** 95% CI: .83 to 2.0 1.03 (.25)** 95% CI: .54 to 1.51 2.13 (.21)** 95% CI: 1.7 to 2.6

3.45**

16.57 (.31)** 95% CI: 15.9 to 17.1 .27 (.05)** 95% CI: .18 to .38 4.5  103 (3.3  103) 95% CI: 2  103 to .01 0 1.50 (.38)** 95% CI: .75 to 2.2 1.52 (.31)** 95% CI: .90 to 2.1 2.57 (.27)** 95% CI: 2.0 to 3.1 0 .02 (.08) 95% CI: .17 to .13 .16 (.06)* 95% CI: .3 to .04 .14 (.05)* 95% CI: .24 to .03 3.45**

.53** .52**

.52** .50**

2.32** 59750.4

2.32** 59741.7

NS*timea ISS*time OSS*time BIOS*time SD of student-intercept SD of student-slope Correlation between random effects of the students SD of the residual 2LLb

BIOS ¼ both in-school and out-of-school sport; CI ¼ confidence interval; ISS ¼ inschool sport; NS ¼ no sport; OSS ¼ out-of-school sport; SE ¼ standard error. *p < .05; **p < .0001. a Reference category. b Computed from a model using full maximum likelihood method.

engagement. Duration of participation in organized activities was associated with enhanced adjustment after controlling for prior adjustment [16]. Therefore, sustained involvement in multicomponent interventions encompassing more than one organized group, including sport, may influence SPSC to a larger extent than sport alone. Future research should evaluate the impact of sport within a multicomponent intervention on various indicators of social and mental well-being during both early and later adolescence. Previous literature has been mixed, showing both differences and similarities between boys and girls and the relationship between sport and self-concept [16,23,33]; therefore, it is difficult to comment on the nonsignificant interaction between sport and sex in this study. It is possible that gender differences will emerge later in adolescence, given that girls tend to withdraw from sport around this time of early adolescence [34]. It is possible that differences might emerge depending on the specific sport context (team vs. individual) or type (soccer vs. gymnastics), which would require a more comprehensive sport assessment. Finally, it remains plausible that the specific relationship between sport and social competence is truly the same for boys and girls. Possible explanations of results Potential mechanisms of action can be hypothesized in consideration of either self-systems or social control theories. Self-systems theory suggests that children internalize significant

7

others’ perceptions of themselves such that they become their self-perceptions [35]. Connell and Wellborn [36] suggest that internalization of others’ perceptions is driven by fundamental psychological needs of competence, autonomy, and relatedness. Participation in sport addresses each of these psychological needs: practices can build athletic and social competence, the choice to participate in sport can be made independently and freely by the adolescent, and sport itself takes place in a social context. Therefore, enhanced perceptions of social competence may reflect internalization of their positive perceptions of their coaches and teammates’ judgment of their social capabilities. This may explain the results in the present study in which participation in BIOS was associated with the largest increase in SPSC because each of the psychological needs may be heightened through exposure to both types of sport. Social control theory posits that individuals that form attachments to social institutions are more likely to engage in behavior that is approved by the institution and less likely to behave in ways that are not accepted [25]. Adolescents may form these bonds with various ecological systems, such as family, school, and peer groups. Sport and physical activity, in general, are viewed as socially approved behaviors among these various systems, such that adolescents are encouraged and subsequently valued for engagement in sports. This theoretically perpetuates a stronger social bond with the social institution, which leads to development of positive self-concept. For example, peers may encourage a fellow student to join the team; once the student joins the team, they feel valued and attached to the team and their peers. Conforming to the socially approved norm, this positive interaction with peers facilitates improvement of perceived social competence and overall self-concept. The results of the present study suggest that involvement in BIOS activities is associated with the largest increase in perceived social competence; therefore, attachments may be formed with both the school and other sport institution in which they are involved. Limitations Notably, this study includes repeated measures of a large sample of participants over a dynamic transition period from late childhood to early adolescence; however, there are limitations that must be acknowledged. Sport participation was selfreported; therefore, responses on this survey may have been affected by recall and social desirability bias such that children may not have accurately reported their level of sport participation. However, the prevalence of sport participation reported in this sample is similar to levels reported by other studies in comparable samples [37], suggesting that the risk of measurement bias in this variable is low. Furthermore, the questionnaire is unable to distinguish between different categories of sport such as team and individual, provides information only on the breadth of sport participation and not the level within each activity, and is unable to capture the quality of the sport experience. However, to date, there is no existing tool that can validly and comprehensively measure the full multidimensional nature of sport [38]. In addition, household income was estimated by linking 2006 census data to postal code information collected via a survey administered to the participants. Although indirect, estimating income through postal code matching has been demonstrated as a valid method [31]. Although the use of repeated measures over 4 years strengthens the evidence of the temporal relationship between sport participation and perceived

8

C. Bedard et al. / Journal of Adolescent Health xxx (2019) 1e8

social competence, establishment of causality cannot occur. It is possible that young adolescents with higher perceptions of social competence self-select into sports. However, the significant relationship over time suggests that sport is still a positive option for these students. Finally, this study is limited by nature of secondary data analysis, and therefore, many of the hypothesized mechanisms of action cannot be tested. Conclusion The results of this study demonstrated that higher levels of participation in sport are associated with small gains in perceptions of social competence in both males and females and that this relationship increases over a 4-year period from late childhood to early adolescence. These results suggest that sport participation may be a small yet important contributor to young adolescents’ perceptions of their social capabilities. Given the relationship between perceived social competence and psychosocial well-being, sport may be an important component piece for schools to focus on to enhance psychosocial health. Future research should evaluate the possible mediating effect of athletic competence, autonomy, relatedness, and institutional attachment, as well as assess distal psychosocial outcomes such as internalizing and externalizing problems, in addition to the proximal outcome of self-concept. Acknowledgments The Physical Health and Activity Study appreciates the commitment by children, parents, and teachers from the District School Board of Niagara. Funding Sources This study was supported by the Canadian Institutes of Health Research (grant no. 66959). References [1] Larson RW. Toward a psychology of positive youth development. Am Psychol 2000;55:170. [2] Hansen DM, Larson RW, Dworkin JB. What adolescents learn in organized youth activities: A survey of self-reported developmental experiences. J Res Adolesc 2003;13:25e55. [3] Eime RM, Young JA, Harvey JT, et al. A systematic review of the psychological and social benefits of participation in sport for children and adolescents: Informing development of a conceptual model of health through sport. Int J Behav Nutr Phys Act 2013;10:98. [4] Holt NL, Kingsley BC, Tink LN, et al. Benefits and challenges associated with sport participation by children and parents from low-income families. Psychol Sport Exerc 2011;12:490e9. [5] Tracy AJ, Erkut S. Gender and race patterns in the pathways from sports participation to self-esteem. Sociol Perspect 2002;45:445e66. [6] Donaldson SJ, Ronan KR. The effects of sports participation on young adolescents’ emotional well-being. Adolescence 2006;41:369e89. [7] Zaharopoulos E, Hodge KP. Self-concept and sport participation. N Z J Psychol 1991;20:12e6. [8] Snyder AR, Martinez JC, Bay RC, et al. Health-related quality of life differs between adolescent athletes and adolescent nonathletes. J Sport Rehabil 2010;19:237e48. [9] Howie LD, Lukacs SL, Pastor PN, et al. Participation in activities outside of school hours in relation to problem behavior and social skills in middle childhood. J Sch Health 2010;80:119e25.

[10] Linver MR, Roth JL, Brooks-Gunn J. Patterns of adolescents’ participation in organized activities: Are sports best when combined with other activities? Dev Psychol 2009;45:354. [11] Lerner RM, Galambos NL. Adolescent development: Challenges and opportunities for research, programs, and policies. Annu Rev Psychol 1998;49:413e46. [12] Dimech AS, Seiler R. Extra-curricular sport participation: A potential buffer against social anxiety symptoms in primary school children. Psychol Sport Exerc 2011;12:347e54. [13] Findlay LC, Coplan RJ. Come out and play: Shyness in childhood and the benefits of organized sports participation. Can J Behav Sci 2008;40: 153e61. [14] Guèvremont A, Findlay L, Kohen D. Organized extracurricular activities: Are in-school and out-of-school activities associated with different outcomes for Canadian youth? J Sch Health 2014;84:317e25. [15] Marsh HW. Age and sex effects in multiple dimensions of self-concept: Preadolescence to early adulthood. J Educ Psychol 1989;81:417. [16] Viau A, Denault A-S, Poulin F. Organized activities during high school and adjustment one year post high school: Identifying social mediators. J Youth Adolesc 2015;44:1638e51. [17] Telama R, Yang X. Decline of physical activity from youth to young adulthood in Finland. Med Sci Sports Exerc 2000;32:1617e22. [18] Anderson S, Messick S. Social competency in young children. Dev Psychol 1974;10:282. [19] Childs HF, Schneider HG, Dula CS. Adolescent adjustment: Maternal depression and social competence. Int J Adolesc Youth 2001;9:175e84. [20] Grisset NI, Norvell NK. Perceived social support, social skills, and quality of relationships in bulimic women. J Consult Clin Psychol 1992;60:293. [21] Fenzel LM. Prospective study of changes in global self-worth and strain during the transition to middle school. J Early Adolesc 2000;20:93e116. [22] McGee R, Stanton WR. Sources of distress among New Zealand adolescents. J Child Psychol Psychiatry 1992;33:999e1010. [23] Cole DA, Maxwell SE, Martin JM, et al. The development of multiple domains of child and adolescent self-concept: A cohort sequential longitudinal design. Child Dev 2001;72:1723e46. [24] Slater A. Gender differences in adolescent sport participation, teasing, selfobjectification and body image concerns. J Adolesc 2011;34:455e63. [25] Hirschi T. Causes of delinquency. Berkeley: University of California Press; 1969. [26] Cairney J, Veldhuizen S. Organized sport and physical activity participation and body mass index in children and youth: A longitudinal study. Prev Med Rep 2017;6:336e8. [27] Harter S. The self-perception profile for children: The perceived competence scale for children (Manual). Denver CO: Univ Denver; 1985:44. [28] Harter S. Self-Perception profile for children: Manual and questionnaires (revision of the self-perception profile for children, 1985). Denver, CO: Univ Denver; 2012. [29] Muris P. The self-perception profile for children: Further evidence for its factor structure, reliability, and validity. Personal Individ Differ 2003;35: 1791e802. [30] Hay JA. Adequacy in and predilection for physical activity in children. Clin J Sport Med 1992;2:192e201. [31] Mustard CA, Derksen S, Berthelot J-M, et al. Assessing ecologic proxies for household income: A comparison of household and neighbourhood level income measures in the study of population health status. Health Place 1999;5:157e71. [32] Skinner RA, Piek JP. Psychosocial implications of poor motor coordination in children and adolescents. Hum Mov Sci 2001;20:73e94. [33] Marsh HW. Extracurricular activities: Beneficial extension of the traditional curriculum or subversion of academic goals? J Educ Psychol 1992; 84:553. [34] Craggs C, Corder K, van Sluijs EMF, et al. Determinants of change in physical activity in children and adolescents: A systematic review. Am J Prev Med 2011;40:645e58. [35] Harter S. The construction of the self: Developmental and sociocultural foundations. 2nd ed. New York, NY: Guilford Publications; 2012. [36] Connell JP, Wellborn JG. Competence, autonomy, and relatedness: A motivational analysis of self-system processes. In: Gunnar MR, Sroufe LA, eds. Minn. Symp. Chid Psychol. Vol 23 Self Process. Dev. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.; 1991:43e77. [37] Sabiston CM, Jewett R, Ashdown-Franks G, et al. Number of years of team and individual sport participation during adolescence and depressive symptoms in early adulthood. J Sport Exerc Psychol 2016;38: 105e10. [38] Clark HJ, Camiré M, Wade TJ, et al. Sport participation and its association with social and psychological factors known to predict substance use and abuse among youth: A scoping review of the literature. Int Rev Sport Exerc Psychol 2015;8:224e50.