Electronic Supplementary Material Figure S1. General Physical Self-Concept Forest Plot

1 Electronic Supplementary Material Figure S1. General physical self-concept forest 2 plot. 3

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

1 23 24 Electronic Supplementary Material Figure S2. Perceived competence forest plot. 25

2 26 27 28 Electronic Supplementary Material Figure S3. Perceived fitness forest plot.

30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52

4 53 Electronic Supplementary Material Figure S4. Perceived appearance forest plot. 54

55 56 57

5 58 Electronic Supplementary Table S1 – Summary of findings from included studies. 59 STUDY SAMPLE DESIGN PSC measure, αa PA measure Analysis Findings PSC and PA Corrected r Correlations

Abbott and N = 1837 Cross-sectional Embodied Image PA recall MANCOVA Individuals involved in either sports NR NR Barber (2011) 1002 girls and 822 boys Scale, 0.81 questionnaire or physical activity held a greater 13 did not specify sex positive body image than those Age range 13-18 considered sedentary. Mean age = 14.7, SD = 1.10 Australia

Abarca-Sos et al N = 1272 Cross-sectional PSDQ, 0.89 Assessment of SEM SC positivity associated with PA. SEM SEM (2013) 590 girls and 682 boys Physical Activity Bivariate Models included changes in PA PA and SC SC, 0.22 Age range 12-15 Levels questionnaire correlations related to age r = 0.22, (p < Mean age = 13.5, SD = 0.7 Accelerometers 0.001) Spain PA and SC r = 0.55, (p < 0.5)

Altintas and N = 803 Cross-sectional CY-PSPP, PA recall MANOVA Females in the high PA group had NR NR Asçi (2008) 383 girls and 420 boys α = DNR questionnaire higher GPSE scores than those in the Age range 11-14 low PA group including SC, PC, Mean age = 13.10, SD = 0.93 STR, GPSW but not BA. Turkey Males in the high PA group had higher GPSE scores than those in the low PA group including SC, PC, STR, GPSW and BA.

6 STUDY SAMPLE DESIGN PSC measure, αa PA measure Analysis Findings PSC and PA Corrected r Correlations

Annesi (2006) 2003 N = 41 Experimental PSCS, 0.70 PA recall Bivariate Increases in PA were significantly 2003 Group 2003 Group 2005 N= 84 (12 Weeks) Exercise Barriers Self- questionnaire correlations associated with increases GPSC over Δ PA and Δ Control N = 40 Efficacy Scale for the 12 weeks. GPSC T1 Age range 9-12 Children, 0.79 r = 0.39, (p < GPSC, 0.47 Mean age = 10.8, SD = 1.1 0.05) T2 USA T2 PA and Δ GPSC, 0.11 GPSC r = 0.09, (p 2005 Group >0.05) T1 2005 Group GPSC, 0.31 Δ PA and Δ T2 GPSC GPSC, 0.25 r = 0.26, (p < 0.05) T2 PA and Δ GPSC r = 0.21, (p > 0.05) Annesi et al N = 269 Experimental PSDQ, 0.81 PA recall Bivariate Increases in PA were significantly PA and BI BI, 0.41 (2008) 159 girls and 110 boys (12 Weeks) questionnaire correlations associated with increases in BI and r = 0.37, (p < PSC, 0.42 Mean age = 10.6, SD = 1.1 PSC over the 12 weeks. 0.001) USA PA and PSC r = 0.38, (p < 0.001) Annesi et al N = 43 Experimental PSDQ, 0.70 PA recall Linear multiple Association between PA and GPSC NR NR (2009) 22 girls and 21 boys (12 Weeks) Exercise Barriers Self- questionnaire regression ( = 0.11) Age range 7-12 efficacy Scale for Muscular STR push (p = 0.47). Mean age = 9, SD = 1.3 Children, 0.72 up test Canada Asci et al (2001) N = 115 Cross-sectional Piers-Harris Weekly activity ANOVA Adolescents in the high PA group NR NR 59 girls and 56 boys Children’s Self- checklist had significantly higher SC than Age range 11-14 Concept Scale those of a low PA group. Mean age girls = 12.69, SD = Reliability of scale 0.74 was 0.87 Mean age boys = 12.86, SD = SPPC Reliability of 0.62 scale was 0.65 Turkey Asci et al (1997) N = 348 boys Cross-sectional PSPP, 0.84 Categorized athletes Bivariate Athletes participating in PA reported NR NR Age range 14-17 and non-athletes. correlations of PSC significantly higher SC but reported Mean age = 15.5 and BI. lower levels of GSW compared to Turkey non-athletes.

Bagoien et al N = 329 Cross-sectional Perceived PA recall SEM SC indicated greater involvement in NR NR (2010) Mean age = 16.5, SD = 0.7 Competence Scale, questionnaire PA 0.58 Norway 0.84 Needs satisfaction significantly influenced effort of PA = 0.36 Baker and N = 149 girls Longitudinal PSPP, 0.84 at age 11 Accelerometers Bivariate Participation in PA was associated Δ in PA and SC SC, 0.21 Davison (2011) Age range 9-13 (24 months) and 0.86 at age 13 correlations with higher SC. SC was a positive r = 0.21, (p < Mean age = 11 predictor of MVPA. 0.05) USA Balaguer et al N = 917 Cross-sectional SPPC, α = DNR PA recall MANOVA For boys and girls, higher levels of NR NR (2012) 460 girls and 457 boys questionnaire sport participation were positively Mean age = 14.46, SD = 1.79 associated to SC, and for boys were Spain also associated with BI.

Self-perceptions were positively associated to frequency levels of sport participation in both sex groups. Barnett et al N = 215 Cross-sectional PSPP, 0.75 PA recall SEM PA was mediated by SC. Direct SEM SC, 0.35 (2011) 111 girls and 104 boys questionnaire Bivariate effects were present between PA and PA and SC Mean age = 16.4, SD = 0.6 correlations SC (β = 0.35). r = 0.35, (p = Australia 0.001)

SEM SC and PA r = 0.23, (p = 0.007) MVPA and SC Direct β = 0.35 Barnett et al 2000 N = 1045 Longitudinal PSPP, 0.80 for girls PA recall SEM SEM SEM SEM (2008) 2006/7 N = 276 (12 Months) and 0.75 for boys questionnaire Bivariate SC and PA PA and SC PA and SC, 0.34 145 girls and 131 boys correlations r = 0.34, (p = < 0.01). SEM included r = 0.34, (p < Age range 7.9-11.9 associations of PA, SC, childhood 0.01) Correlation Mean age = 10.1 locomotor skill and childhood object Girls Australia control skills. Positive SC is a PA and SC girls SC, 0.41 predictor of PA. r = 0.37, (p = 0.01) Boys PA and SC boys SC, 0.29 r = 0.25, (p = 0.01)

Bevans et al N = 2018 Cross-sectional Single-item perceived PA recall SEM Physical activity was directly SEM SEM (2010) Age range 10-14 competence in PE questionnaire, 0.83 affected by student engagement and Mean age = 12.2 questionnaire, α = perceived competence in PE and PA and SC PA and SC, 0.20 USA DNR indirectly affected by students’ body r = 0.20, (p <

image through its association with 0.0001) PA and BI, Healthy Pathways PE engagement. PA and BI Child-Report Body 0.05 Image r = 0.05, (p > Scale, 0.70 for BI 0.05)

Biddle and N = 72 Cross-sectional PSPP-C, Heart rate monitors ANOVA Active girls were characterised by NR NR Armstrong 35 girls and 37 boys α = DNR higher (1992) Age range 11-12 scores on perceptions of BI, as well Mean age = 12.2, SD = 0.31 as England PSW and GSE. For boys it was not possible to discriminate active from less active on the basis of their scores on the psychological inventories Biddle and N = 147 Cross-sectional PSPP, 0.51 LTEQ SEM PA was predicted by SC SEM SEM Goudas (1996) Age range 13-14 Bivariate SC to PA SC to PA, Mean age = 13.5 correlations 0.18 England r = 0.18, (p < 0.1)

Correlations PA and SC r = 0.30, (p < 0.01) Biddle and N= 516 girls Cross-sectional PSPP-C, PA recall Bivariate PA was significantly correlated to PA and SC SC, -0.25, Wang (2003) Age range 11–16 SC, 0.78, questionnaire correlations GPSC and PC. BI, 0.18, Mean age = 13.69, SD = 0.93 BI, 0.87, Models included associations of BA r = -0.22, (p < STR, 0.11 England STR, 0.81 and GPSW. 0.001) PA and BI r = 0.17, (p < 0.001) PA and STR r = 0.10, (p <

0.005) Bohnert and N = 76 Experimental Rosenberg Self- PA recall ANOVA Girls reported higher levels of body NR NR Ward (2013) 76 girls and 57 boys (30 weeks) Esteem Scale, 0.83 questionnaire dissatisfaction at the conclusion of Age range 8-12 programming.

Mean age = 9.3, SD = 1.00 USA Burgess et al N = 50 Experimental BAQ, LTEQ ANOVA PA resulted in improvements to SC, NR NR (2006) Age range 13-14 (6 Weeks) BI, 0.80, PC, BI, STR and PSW. Mean age =13.5, SD = 0.3 STR, 0.86 UK CY-PSPP, PSW, 0.91, SC, 0.88, PC, 0.89, BI, 0.94, STR, 0.77 Carroll and N = 922 Cross-sectional Self-perceived PA recall MANOVA Individuals whom perceived NR NR Loumidis (2001) 454 girls and 468 boys competence in PE questionnaire themselves as more competent in PE Age range 10-11 scale, 0.78 participate in more PA and higher Britain levels of intensity than those who perceived themselves to be less competent. Chen et al N = 883 Cross-sectional Multidimensional Taiwan National Bivariate PA was positively related with body Girls Girls BI, (2010) 431 girls and 452 boys Body-Self Relations Physical Activity correlations dissatisfaction for girls, but not for PA and BD -0.14 Age range 12-16 Questionnaire – Survey boys. r = 0.19, (p < Taiwan Appearance 0.01) Boys BI, 0.02 Evaluation, 0.74 PA and BI r = -0.12, (p > 0.05)

Boys PA and BD r = -0.02, (p > 0.05) PA and BI r = 0.02, (p > 0.05) Cheng et al N = 206 girls Cross-sectional Body image and PA recall Bivariate PA participation was significantly PA and BI BI, 0.20 (2003) Age range 11-18 psychological questionnaire correlations correlated to BI. r = 0.17, (p = Hong Kong questionnaire, 0.02) α = DNR

Craft et al N = 46 girls Cross-sectional PSPP-A, Heart-rate monitor Bivariate SC is significantly correlated to PA. PA and SC SC, 0.35, (2003) 12.0, SD = 0.6 years PSW, 0.73, transmitter belt correlations r = 0.32, (p < BA, 0.29, USA BI, 0.80, 0.05) GPSW, 0.20 SC, 0.83 PA and BA r = 0.26, (p > 0.05) PA and GPSW r = 0.17, (p > 0.05) Crocker et al N = 631 girls Longitudinal PSPP, PAQ-A Bivariate PA change was associated with T1 PSW T1 (2003) Age range 15-16 (24 months) α = DNR correlations change in all r = 0.29, (p < PSW, 0.35, Canada self-perceptions except BMI 0.05) T1 BA, 0.14, T1 BA T1 PC, 0.61, r = 0.12, (p < T1 SC, 0.55, 0.05) T1 STR, 0.38 T1 PC r = 0.51, (p < T2 0.05) PSW, 0.48, T1 SC T2 BA, 0.20, r = 0.46, (p < T2 PC, 0.69, 0.05) T2 SC, 0.63, T1 STR T2 STR, 0.42 r = 0.32, (p < 0.05)

T2 PSW r = 0.40, (p < 0.05) T2 BA r = 0.17, (p < 0.05) T2 PC r = 0.58, (p < 0.05) T2 SC r = 0.53, (p < 0.05) T2 STR r = 0.35, (p < 0.05) Crocker et al N = 501 girls Longitudinal PSPP, PAQ-A Bivariate Correlations at 3 intervals indicated Cross-sectional T1 (2006) Age range 14-15 (1st year) (24 months) STR 0.87, correlations that all physical self-perceptions and T1 SC, 0.50, Age range 16–17 (3rd year) T1 = baseline SC 0.88, global self-esteem scores were PA and SPA BA, 0.13, Canada T2 = approximately BI, 0.87, significantly correlated with PA; r = -0.08, (p > STR, 0.35,

one year after PSW, 0.87, GSE, with PC being the dominant 0.05) GPSW, 0.32 baseline 0.79, correlate. PA and GPSW T3 = final set SPAS, 0.85 PC (β = 0.48), SC (β = 0.19), and r = 0.30, (p < T2 BA (β = −0.14) were significant 0.05) SC, 0.55, individual predictors of PA. PA and BA BA, 0.15, r = 0.12, (p < STR, 0.39, 0.05) GPSW, 0.42 PA and PC r = 0.53, (p < T3 0.05) SC, 0.54, PA and SC BA, 0.19, r = 0.47, (p < STR, 0.41, 0.05) GPSW, 0.40 PA and STR r = 0.33, (p < LS 0.05) SC, 0.28, BA, 0.11, T2 STR, 0.24, PA and SPA GPSW, 0.19 r = -0.09, (p < 0.05) PA and GPSW r = 0.39, (p < 0.05) PA and BA r = 0.14, (p < 0.05) PA and PC r = 0.57, (p < 0.05) PA and SC r = 0.52, (p < 0.05) PA and STR r = 0.36, (p < 0.05) T3 PA and SPA r = -0.16, (p < 0.05) PA and GPSW r = 0.37, (p < 0.05) PA and BA

r = 0.18, (p < 0.05) PA and PC r = 0.55, (p < 0.05) PA and SC r = 0.51, (p < 0.05) PA and STR r = 0.38, (p < 0.05)

Longitudinal PA and SPA r = -0.08, (p > 0.05) PA and PSW r = 0.18, (p < 0.05) PA and BA r = 0.10, (p < 0.05) PA and PC r = 0.34, (p < 0.05) PA and SC r = 0.26, (p < 0.05) PA and STR r = 0.22, (p < 0.05)

Crocker et al N = 466 Cross-sectional PSPP, PAQC SEM All physical self-perceptions were Girls Girls (2000) 246 girls and 220 boys α = DNR Bivariate statistically significantly (p < 0.05) PA and PC SC, 0.55, Age range 10-14 correlations correlated with PA among both girls r = 0.47, (p < BI, 0.32, Canada (r = 0.26-0.47) and boys (r = 0.28- 0.05) STR, 0.43, 0.47). All PSPP sub-domains were PA and SC GPSW, 0.45 moderately correlated in boys (r = r = 0.46, (p < 0.48-0.68) and girls (r = 0.42-0.67). 0.05) Boys PA and BA SC, 0.55, SEM Girls r = 0.27, (p < BI, 0.33, PA and SC 0.05) STR, 0.42, r = 0.28, (p > 0.05) PA and STR GPSW, 0.47 PA and STR r = 0.36, (p < r = 0.05, (p > 0.05) 0.05) PA and BI PA and GPSW r = 0.10, (p > 0.05) r = 0.38, (p < 0.05 SEM Boys PA and SC Boys r = 0.37, (p > 0.05) PA and PC PA and STR r = 0.47, (p < r = 0.07, (p > 0.05) 0.05) PA and BI PA and SC r = 0.24, (p > 0.05) r = 0.46, (p < 0.05) PA and BA r = 0.28, (p < 0.05) PA and STR r = 0.35, (p < 0.05) PA and GPSW r = 0.39, (p < 0.05) Cuddihy et al N = 297 girls Cross-sectional CY-PSPP, Pedometers ANOVA All means for NR NR (2006) Age range 13-15 α = DNR self-perceptions are higher for Australia higher physical activity levels except Global Self- Worth.

Cumming et al N = 407 girls Cross-sectional CY-PSPP, PAQ-A SEM PSW, positively predicted (SEM) SEM (2011) Age range 11-15 α = 0.84 (SC) Bivariate involvement in physical activity PSW and PA SC, 0.51, Mean age =13.2, SD = 1.0 α = 0.87 (PC) correlations (β = 0.44, p < 0.001). r = 0.43 (p < BI, 0.25, England α = 0.91 (BI) 0.05) STR, 0.32

α = 0.87 (STR) PA and SC α = 0.87 (PSW) r = 0.47, (p < 0.001) PA and PC r = 0.49, (p < 0.001) PA and BI r = 0.24, (p < 0.001) PA and STR r = 0.30, (p < 0.001) PA and PSW r = 0.36, (p < 0.001) Daley (2002) N = 1230 Cross-sectional CY-PSPP, PA questionnaire from ANOVA Univariate analyses revealed that NR NR 601 girls and 629 boys α = DNR the ‘Young People children who indicated that they Age range 14-15 and Sport’ survey participated in extra-curricular PA England reported significantly higher scores for BA, (F = 11.26, p < 0.01) and GPSW (F = 13.55, p = 0.01). DeBate et al N = 1034 Experimental Rosenberg Self- Youth Risk T-tests Paired sample t-tests revealed NR NR (2009) Age range 8-15 (12 weeks) Esteem Scale, Behavioural statistically significant differences in USA α = 0.79 Survey most variables from pre- to post Schematic PA recall intervention including self-esteem (p Figural Scale, questionnaire < 0.001), body size satisfaction (p < α = 0.62 Pre 0.001), and the number of VPA days Intervention, within the last week (p < 0.001). α = 0.69 Post Intervention In the younger age group, statistically significant differences from pre- to post intervention were found for self-esteem (p < 0.001), body size satisfaction (p < 0.001), and PA frequency (p < 0.001). Among the older girls, Statistically significant differences were also observed for self-esteem (p < 0.001), body size satisfaction (p < 0.001), PA frequency (p < 0.001) as well as commitment to PA (p = 0.003).

The current study revealed a Cronbach alpha of 0.79 for both pre- and post-intervention. Dishman et al N = 1250 girls Cross-sectional PSDQ, 3-Day Physical SEM 3DPAR had significant, direct SEM SEM (2006) Mean age =17.66, SD = 0.61 α = 0.80 Activity Recall Bivariate relations with the subscales of PA and STR PA and STR, USA correlations Strength, and Sports Competence, r = 0.14, (p > 0.14, 0.05) PA and SC, Significant path coefficients were PA and SC 0.10, also found from Sport participation r = 0.10, (p > PA and BI, 0.07, to each of the subscales of physical 0.05) PA and GSE, self-concept. PA and BI 0.13 r = 0.07, (p > Results suggest physical self- 0.05) concept mediates the relations of PA and GSE physical activity and sport r = 0.13, (p > participation with self-esteem. 0.05)

PA and GPSC r = 0.41, (p > 0.05) PA and STR r = 0.48, (p > 0.05) PA and BI r = 0.04, (p > 0.05) PA and SC r = 0.50, (p > 0.05) Douthitt (1994) N = 110 Cross sectional PSPP, The Habitual PA Bivariate Girls SC, GSW and BI were Girls PA and SC Girls 33 girls and 77 boys α = DNR Questionnaire correlations predictive of females exercise r = -0.51, (p = SC, -0.61, Age not determined (Grade 9- adherence. 0.003) GSW, -0.51, 11, age range 14-16) Girls PA and BI, -0.10 USA GSW r = -0.43, (p = 0.02) Girls PA and BI r = -0.09, (p = 0.007) Duncan et al N = 277 Cross-sectional Body Esteem Scale PA recall Bivariate Results indicated no significant Girls Girls, BE, (2004) 111 girls and 166 boys for children, questionnaire correlations associations between body image PA and Body -0.19 Age range 11-14 α = DNR and PA (p > 0.05). Esteem Boys, BE, 0.06 Mean age = 12.5, SD =0.8 r = -0.16, (p > United Kingdom 0.05)

Boys PA and Body Esteem r = 0.05, (p > 0.05) Duncan et al N = 276 Cross-sectional Figure rating scale, PA recall Bivariate Associations were evident between Girls Girls, 0.19 (2006) 110 girls and 166 boys α = DNR questionnaire correlations average daily energy expenditure PA and current Age range 11-14 and children’s perceived current body shape Boys, 0.56 Britain body shape for all children (r = 0.09) r = 0.12, (p > (p > 0.05). 0.05)

Boys reported a significant Boys association with PA and BA. PA and current body shape r = 0.47, (p > 0.05) Dunton et al N = 89 girls Cross-sectional PSDQ, 2-Day Physical Bivariate Check PAPER PA and BI BI, -0.11; (2003) Age range 14-17 α = DNR Activity Recall correlations r = -0.09, (p > STR, 0.35; Mean age =15.03, SD = 0.79 0.05) GPSC, -0.25 USA PA and END r = -0.02, (p > 0.05) PA and STR r = 0.29, (p < 0.01) PA and CRD r = 0.07, (p > 0.05) PA and FLX r = -0.02, (p > 0.05) PA and GPSC r = -0.21, (p > 0.05) PA and SE r = 0.09, (p > 0.05) Dunton et al N = 103 girls Cross-sectional PSDQ, α = DNR PA recall Bivariate Pearson’s correlations showed that PA and SC SC, 0.4; (2006) Age range 14 -17 questionnaire correlations vigorous r = 0.36, (p < BI, 0.54; Mean age =15.24, SD = 0.81 physical activity was positively 0.001) STR, 0.54; USA associated with scores on most of PA and BI PSW, -0.47 the PSDQ scales r = 0.43, (p < (p < 0.005). 0.001)

PA and STR r = 0.45, (p < 0.001) PA and FLX r = 0.23, (p > 0.05) PA and END r = 0.37, (p < 0.001) PA and PSW r = -0.33, (p < 0.005) PA and SE r = 0.47, (p < 0.001) Eccles and T1 Cross-sectional Self-Concept of PA recall SEM PC was positively associated with (SEM) PA and (SEM) PA and Harold (1991) N = 2700 Ability questionnaire, questionnaire Bivariate free time involvement in sport in PSC PSC T2 0.70 correlations both girls and boys. r = -0.11(p = r = -0.11(p = N = 875 0.46) 0.46) USA Girls PA and PC r = 0.47, (p < 0.001) Boys PA and PC r = 0.44, (p < 0.001) Findlay and N = 201 Longitudinal PSDQ Sport Participation MANOVA Children who participated in sport NR NR Coplan (2008) 105 girls and 96 boys (12 months) T1 and T2 Information Sheet reported more competence, physical Age range T1 8.9-11.8 respectively. ability, physical appearance and self- Mean age =10.01, SD = 0.6 BI α = 0.89 and α = esteem Canada 0.91) Gao (2008) N=307 Cross-sectional The self-perceived Pedometers Multiple regression PA was positively associated with NR NR 158 girls and 149 boys competence in PC ( = 0.23) (p < 0.05) Age range 12-15 physical education USA scale α = 0.78

Gaspar et al N = 234 Cross-sectional Collins’ Child Habitual PA Logistic regression PA exerted a protective effect on NR NR (2011) 113 girls and 121 boys Figure Drawings scale questionnaire satisfaction with BI Age range 10.3-17.7 α = 0.71 Portugal

Gillison et al N = 310 Cross-sectional Body Change LTEQ Bivariate BI and SC were positively PA and BI BI, 0.15; (2011) 153 girls and 157 boys Inventory correlations associated with the maintenance of r = 0.14, (p < SC, 0.30 Age range 13.88-15.64 α = 0.82 PA. 0.01) Mean age =14.93, SD = 0.30 PA and SC England r = 0.25, (p < 0.001)

Gísladóttir et al N = 10987 Cross-sectional Youth in Iceland Offer Self-Image Regression analysis PA was associated with higher BI NR NR (2013) 5424 girls, 5369 boys and 194 survey Questionnaire and PC did not specify sex. α = 0.83 Age range 14-16 Iceland

Goldfield et al N = 1259 Experimental Body esteem scale for Godin leisure-time Partial correlations Vigorous PA was significantly Girls Boys (2011) 746 girls and 513 boys adolescents and adults exercise questionnaire correlated with the external Moderate PA BI, 0.22 Age range 12-18 α = 0.86 attribution subscale of the body and BI Canada esteem scale for adolescents in r = 0.00, (p > males (r = 0.20) (p < 0.001) but not 0.05) in females (r = 0.00) (p > 0.05). Vigorous PA Vigorous PA was the strongest and BI correlate of positive body image r = 0.00, (p > findings in the overall sample. 0.05)

Boys Moderate PA and BI r = 0.06, (p > 0.05) Vigorous PA and BI r = 0.20, (p < 0.001) Guinn et al N = 254 girls Cross-sectional Dusek’s Abbreviated PA recall Bivariate BI was statistically significantly PA and BI r = BI, 0.30 (1997) Age range 13-15 form of Second- questionnaire correlations associated with PA involvement r = 0.27, (p < 0.001) Mean age 14.4 Jourard Body Cathexis 0.27, (p < 0.001) Mexico Scale α = 0.83

Hagger et al N = 252 Cross-sectional PSPP-C PA recall MANOVA Both groups were subjected to NR NR (1998) 134 girls and 118 boys α = DNR questionnaire discriminant analysis to examine the Age range 13-14 ability of PSPP-C sub-domains to

British discriminate between high and low active children. N = 240 127 girls and 113 boys Age range 13-14 Russian Haugen et al T1 = 2005 Cross-sectional SPAA PA recall Bivariate Results indicated PA predicted SC in Girls Girls SC, 0.27; (2013) N = 1207 SC α = 0.82 questionnaire correlations both sexes, but not BI. PA and SC Girls BI, 0.07 T2 = 2008 BI α = 0.88 r = 0.24, (p =< N = 632 PA was strongly associated with SC, 0.01) Boys SC, 0.28; Total N = 1839 lesser with BI in both girls and boys. PA and BI Boys SC, 0.01 889 girls and 950 boys. r = 0.07, (p => Age 15 years An inverse association was present 0.05) Norway among girls and boys when examined though the direct fitness Boys outcomes effects on PA. PA and SC r = 0.25, (p =< 0.01) PA and BI r = 0.01, (p => 0.05)

Direct girls PA and SC r = 0.007, (p = > 0.05) PA and BI r = -0.002, (p = > 0.05)

Direct boys PA and SC r = 0.073, (p = < 0.01) PA and BI r = -0.015, (p = < 0.01) Inchley et al N = 641 Longitudinal Children’s Physical PAQ-C Multivariable High self-efficacy and high NR NR (2011) 312 girls and 329 boys (5 years) Activity Self- logistic regression perceived competence were Age range 11-15 Efficacy Scale Chi-squared positively associated with physical Scotland Rosenberg Self- associations activity among both sexes across all Esteem Scale time points. High self-esteem was α = DNR associated with physical activity among P7 and S4 girls only.

Physical self-worth was positively associated with physical activity among girls in all three years and among boys in S2 and S4.

Ingledew and N = 193 (Younger) Cross-sectional Figure rating scale PA recall Bivariate PA did not correlate significantly PA and BI Girls, 0.07 Sullivan (2002) 50 girls and 49 boys. BI α = 0.85 questionnaire correlations with BI in any sex /age group. (Younger girls) Age range 11-13 r = 0.06, (p > Boys, 0.13; N = 94 (Older) 0.05) Older Boys, 49 girls and 45 boys PA and BI -0.15 Age range 17-19 (Older girls) Germany r = 0.00, (p > 0.05) PA and BI (Younger boys) r = 0.12, (p > 0.05) PA and BI (Older boys) r = -0.14, (p > 0.05) Jaakkola et al N = 237 Longitudinal study PSPP PA recall SEM Perceived competence was a (SEM) PA and Girls SC, 0.35 (2013) 101 girls and 136 boys (3 years) SC α = 0.89 questionnaire Bivariate statistically significant mediator of SC T1 Age 13 correlations self-reported PA of both sexes. r = 0.43(.08) (p Boys SC, 0.39 T12 Age 14 < 0.05) T3 Age 15 Finland Girls PA and SC r = 0.33, (p < 0.001) Boys PA and SC r = 0.37, (p < 0.001) Jackson et al N = 244 girls Cross-sectional CY-PSPP PAQ-A SEM SC, PC, BI, STR and GSW were (SEM) PSW and SC, 0.52; (2013) Age range 11-14 α = 0.84 α = 0.75 Bivariate statically significant in positively PA PC, 0.58; Mean age =12.8, SD =0.9 correlations predicted involvement r =0.45(.07), (p BI, 0.22; England in PA < 0.05) STR, 0.44; GSW, 0.42 PA and SC r =0.48, (p < 0.001) PA and PC r =0.50, (p <

0.001) PA and BI r =0.20, (p < 0.01) PA and STR r =0.40, (p < 0.001) PA and GSW r =0.37, (p < 0.001) Jensen and N = 376 Cross-sectional Figure rating scale PA recall Hierarchical Girls who reported high BI engaged NR NR Steele (2009) 173 girls and 203 boys α = DNR questionnaire Multiple Regression in more PA than girls with lower BI. Mean age =10.8, SD =0.65 Analyses No significant associations were USA found between boys BI and self- reported PA. Kalaja et al N = 316 Cross-sectional Sport Competence PA recall Bivariate Negative association between PA PA and SC SC, -0.35 (2010) 162 girls and 154 boys Scale questionnaire correlations and SC r = -0.30, (p < Age 13 α = 0.73 0.01) Finland Knowles et al N = 150 girls Longitudinal CY-PSPP PAQ-C α = 0.85 Bivariate PA change was associated with PA and SC SC, 0.34; (2009) Mean age =12.79, SD = 0.31 (12 months) PC α = 0.81 correlations change in all of the Children and r = 0.31, (p < BI, 0.20; Scotland STR α = 0.82 Youth’s Physical Self-Perception 0.05) PC, 0.34; PSW α = 0.85 Profile (CY-PSPP) subscales. PA and BI STR, 0.22; r = 0.17, (p > PSW, 0.30 0.05) PA and PC r =0.29, (p < 0.05) PA and STR r = 0.20, (p < 0.05) PA and PSW r = 0.28, (p < 0.05) Kololo et al N = 2277 Cross-sectional Body Image Sub-scale Moderate-to-Vigorous Logistic regression A negative self-assessment of body NR NR (2012) 1191 girls and 1086 boys from the Body Physical Activity image was associated with an Age 15 years Investment Scale scale increased risk of insufficient PA Poland α = 0.85 (OR = 1.29; CI (OR): 1.02-1.63). Positive body image reduced the risk of having insufficient PA (OR = 0.64; CI (OR): 0.52-0.79). Lau et al (2004) N = 100 Cross-Sectional PSPP PA recall Bivariate SC was positively associated with PA and SC SC, 0.30 45 girls and 55 boys α = 0.86 questionnaire correlations sport participation. r = 0.28, (p < Age range 12-13 0.05)

Mean age = 12.54, SD =0.44 England Lau et al (2006) N = 469 Cross-Sectional PSPP PA recall Bivariate PA was significantly associated with PA and SC SC, 0.26 231 girls and 238 boys SC α = 0.76 questionnaire correlations SC. r = 0.23, (p < Age range 12-13 0.01) England

Loucaides et al N = 256 Cross-Sectional Self-efficacy PA recall Bivariate PA was significantly associated with PA and SC SC Summer, (2004) Age range 11-12 questionnaire questionnaire correlations SC I both summer and winter. Summer 0.33; Cyprus SC α = 0.79 r = 0.29, (p < SC Winter , 0.30 0.001) PA and SC Winter r = 0.27, (p < 0.001) Lubans et al N = 1518 girls Cross-sectional Physical activity self- Accelerometers SEM All the physical self-concept (SEM) PA and SC, 0.24; (2011) Mean age = 13.4, SD = 0.4 efficacy questionnaire Bivariate subscales (i.e. SC, BA, PC and STR) SE PC, 0.25; Australia α = DNR correlations were all statistically significant and r = 0.11, (p < BI, 0.14; associated with PA. 0.05) STR, 0.18 (SEM) PA and PSW was significantly associated PSW with PA in the SEM. The model also r = 0.07, (p < included PA and SE, enjoyment of 0.05) PA; school PA, social support and the use of PA behavioural strategies. PA and PSW r = 0.17, (p < 0.01) PA and SC r = 0.20, (p < 0.01) PA and PC r = 0.21, (p < 0.01) PA and BA r = 0.12, (p < 0.01) PA and STR r = 0.15, (p < 0.01) Luszczynska N = 551 Longitudinal Perceived Godin Leisure-Time Bivariate SC at baseline predicted PA (2- T1 – PA and SC SC, 0.28 and Abraham 318 girls and 233 boys (14 months) Competence Scale of Exercise correlations month follow-up) among young girls r = 0.23, (p < (2012) Age range 15-18 Intrinsic Motivation Questionnaire and boys. 0.001) Mean age = 16.43, SD = 0.60 Inventory Poland T1 α = 0.65 Higher physical competence (T1)

T3 α = 0.67 predicted PA (T2), regardless of sexes. Multidimensional Body-Self Relations Questionnaire T1 α = 0.84 T3 α = 0.83

Malete (2004) N = 903 Cross-sectional SPAA PA recall MANOVA PC was not associated with PA. NR NR 492 girls and 411 boys α = 0.40 questionnaire Age Range 13-18 Mean age = 16, SD = 1.10 Africa Malete et al N = 1052 Cross-sectional PSPP PA recall Bivariate PSP subscales were not associated PA and PSW r = BA, -0.12; (2008) 614 girls, 426 boys and 12 did PSW α = 0.77 questionnaire correlations with patterns of involvement in -0.13, (p < 0.01) PSW, -0.15 not specify sex BI α = 0.56 sport. PSW was negatively PA and BA r = Age Range 12-19 SC α = 0.50 associated with participating in -0.09, (p < 0.05) Mean age = 14.61, SD = 1.17 sport. PC was not associated with PA and PC r = Jamaica participating in sport. -0.00, (p > 0.05)

Markland and N = 98 Cross-sectional Figure Rating Scale LTEQ Bivariate Check with DL about two PA and Body PA and BD, Ingledew (2007) 48 girls and 50 boys α = 0.91 correlations calculations Discrepancies -0.08 Mean age girls = 16.88, SD r = -0.08, (p > =0.79 0.05) PA and Mean age boys = 16.90, SD introjected =0.74 PA and regulation, 0.41 UK introjected regulation r = 0.37, (p < 0.001) Marsh (1996) N = 192 Cross-sectional PSDQ PA self-report Bivariate PSDQ subscales were significantly Times/typical STR, 0.43; 79 girls and 113 boys BI α = 0.91 (hours/typical week) correlations associated with PA in girls and boys. week SC, 0.39; Age range 13-15 STR α = 0.92 Linear regression PA and STR r = BA, 0.13; Australia END α = 0.92 0.41, (p < 0.05) GPSW, 0.25 FLX α = 0.90 PA and SC r = CRD α = 0.91 0.38, (p < 0.05) SC α = 0.94 PA and BA r =

GPSW α = 0.91 0.12, (p < 0.05) PA and GPSW r = 0.24, (p < 0.05) Marsh et al N = 2786 Longitudinal PSPP PA recall Multilevel Reciprocal effects model results NR NR (2006) 1393 girls and 1393 boys (6 months) α = 0.81 questionnaire modelling show there were was statistically Greece significant effects of T1 physical self-concept on T2 exercise behaviour and of T1 exercise behaviour on T2 physical self- concept.

Martin et al N = 171 Cross-sectional PSDQ PA recall Bivariate PA was positively and significantly PA and BI BI, 0.19; (2006) 75 girls and 96 boys α = 0.92 questionnaire correlations correlated with flow and with the r = 0.18, (p < STR, 0.22; Mean age = 13.57, SD = 1.38 appearance, strength, endurance, 0.05) SC, 0.27 Australia flexibility, and general physical PA and STR subscales of physical self-concept. r = 0.21, (p < PA level was not correlated with 0.05) health, coordination, or global PA and END esteem. r = 0.26, (p < 0.01) PA and FLEX r = 0.21, (p < 0.01) PA and CRD r = 0.16, (p > 0.05) PA and SC r = 0.26, (p < 0.01) Monteiro et al N = 234 Cross-sectional Collins’ Child Figure Baecke questionnaire Logistic regression High levels of PA were associated NR NR (2011) 113 girls and 121 boys Drawings scale Habitual PA index with a protective effect on negative Age range 10-17 α = DNR BI. Mean age = 10.3, SD = 17.7 Portugal

Moreno and N = 2330 Cross-sectional PSPP PA recall MANOVA Individuals whom participated in PA NR NR Cervello (2005) 1130 girls and 1200 boys SC α = 0.80 questionnaire once a week or less had lower scores Mean age = 14.8, SD =0.91 BI α = 0.76 in SC, PC and STR than those that Spain PC α = 0.75 participated in PA more than 3 times STR α = 0.60 a week.

Moreno-Murcia N = 565 Cross-sectional PSPP PA self-report ANOVA Exercisers had higher BA, SC and NR NR (2005) 306 girls and 259 boys SC α = 0.78 PC than non-exercisers. Students

Age range 12-16 BI α = 0.79 who participated in PA outside PE Mean age = 14.5, SD = 0.49 PC α = 0.76 had significantly higher PC. Spain STR α = 0.67 For all subscales (p < 0.001).

Moreno-Murcia N = 472 Cross-sectional PSPP Habitual PA Bivariate In boys, SC ( = 0.77) and BA ( = PA and SC r = SC, 0.67; et al (2011) 213 Girls and 259 boys PC α = 0.88 questionnaire correlations 0.15) were significantly associated 0.63, (p < 0.01) BA, 0.23 Age range 16-20 BI α = 0.83 SEM with PA. PA and BA r = Mean age = 17.37, SD =0.95 In girls, SC ( = 0.70) was 0.21, (p < 0.01) Spain positively and BA ( = -0.13) negatively associated with PA. The SEM also included PA intention and tobacco and alcohol consumption. Morgan et al N = 97 Longitudinal CY-PSPP Pedometers Bivariate Physical self-perception measures T1 PA and T1 T1 (2008) Mean age = 13.5, SD =0.71 (27 months) SC α = 0.80 correlations were significantly related to changes cognitions SC, 0.47; USA PC α = 0.83 Multiple regression in steps/day over a 27-month period. SC r = 0.42, (p PC, 0.40; BI α = 0.88 < 0.01) BA, 0.24; STR α = 0.80 SC emerged the most important PC r = 0.36, (p STR, 0.28; PSW α = 0.88 predictor and inversely related to PA < 0.01) PSW, 0.41 change over a 27-month period. BA r = 0.22, (p <0.05) T2 STR r = 0.25, (p SC, 0.08; <0.05) T2 PC, 0.26; PSW r = 0.38, T2 BA, 0.19; (p <0.01) T2 STR, 0.10; T2 PSW, 0.30 T1 PA and T2 cognitions SC r = 0.07, (p > 0.05) PC r = 0.24, (p < 0.05) BA r = 0.18, (p > 0.05) STR r = 0.09, (p >0.05) PSW r = 0.28, (p <0.01) Morgan et al N = 120 Longitudinal Health behaviour PA recall ANOVA Physical self-perception was NR NR (2003) 64 girls and 56 boys (6 months) survey questionnaire significantly associated with PA T1 mean age = 11.6, SD α = DNR Accelerometers levels. =0.06 T2 Mean age = 12.2, SD

=0.06 USA

Neumark- N = 4746 Cross-sectional Body shape Modified leisure time Multiple and Boys with lower BA reported NR NR Sztainer et al Age range 11-18 satisfaction scale exercise questionnaire logistic regression significantly less PA. Girl’s trends (2004) Mean age = 14.9 α = 0.92 Youth risk behaviour were similar, but associations were USA survey not statistically significant.

Neumark- N = 2516 Longitudinal Body Shape LTEQ Compared mean lower body satisfaction predicted NR NR Sztainer et al 1386 girls and 2516 boys (5 years) Satisfaction Scale and percentages of lower levels of (2006) T1 Younger Mean age = 12.8, T1 and T2 MVPA SD =0.8 T2 Younger Mean age = 17.2, SD =0.6 T1 Older Mean age = 15.8, SD =0.8 T2 Older Mean age = 20.4, SD =0.8 USA Niven et al N = 162 girls Longitudinal Social Physique PAQ-C Bivariate Social physique anxiety was not T1 T1 Girls BA, (2009) Mean age = 11.8, SD =0.3 (6 months) Anxiety Scale correlations related to current or future PA. PA and BA -0.11; UK Internal reliability α = r = -0.10, (p > T2 Girls BA, 0.78 0.05). -0.16

Motivation for T2 Physical Activity PA and BA Measure r = -0.14, (p > SC α = 0.87 0.05). BI α = 0.88

Niven et al N = 208 girls Cross-sectional PSPP-C PAQ-C Bivariate Positive association exists between PA and SC r = SC, 0.49; (2007) Mean age = 11.83, SD =0.39 α = DNR correlations PA and (SC, PC, STR, BA and 0.41, (p < 0.01) PC, 0.50; UK GPSW). PA and PC r = BA, 0.25; 0.42, (p < 0.01) STR, 0.42; PA and BA GPSW, 0.33 r = 0.21, (p < 0.01) PA and STR r = 0.35, (p <

0.01) PA and GPSW r = 0.28, (p < 0.01) Papaioannou N = 1393 Cross-sectional PSPP-C PA self-report ANOVA Students who were not involved in NR NR (1997) Junior school mean age = 13, α = DNR out-of-school sport activities had SD =0.05 lower scores on SC. Senior school mean age = 16, SD =0.05 Greece Papaioannou et N = 882 Longitudinal PSPP PA self-report SEM SC was statistically significant to PA and SC SC, 0.18 al (2006) 553 girls and 329 boys (14 months) α = DNR PA r = 0.15, (p < Age range 11-18 0.05) Greece Task and Ego Orientation in Physical Education Questionnaire α = DNR Paxton et al N = 63 Cross-sectional Perceived physical PAQ-C Bivariate Bivariate correlations were PA and SC r = SC, 0.40; (2004) Age range 9-14 competence scale correlations significant between PA and SC, PA 0.34, (p < 0.01) BI, 0.53 Mean age = 11.5, SD = 1.77 Scale reliability α = and BI. PA and BI r = USA 0.71 0.45, (p < 0.01)

Planinšec and N = 364 Cross-sectional Children’s Physical PA self-report ANOVA The high active PA group scored NR NR Fošnarič (2005) 185 girls and 179 boys Self Concept Scale significantly higher on the GPSC, Mean age = 6.4, SD =0.3 BA, SP) than the low active group. Slovenia

Planinsec et al N = 528 Cross-sectional PSPP-C PA recall ANOVA Children from the high activity NR NR (2004) 277 girls and 251 boys α = DNR questionnaire group achieved statistically Age range 5.7 – 8.4 significant and higher results in Mean age = 6.8, SD =0.05 GPSC including BI. Slovenia

Plotnikoff et al N = 2688 Cross-sectional Body weight LTEQ Bivariate PA was negatively and significantly Girls PA and BI Girls BI, (2007) 1446 girls and 1242 boys perceptions correlations correlated to BI for girls and boys. r = -.07, (p < -0.08 Mean age = 15.6 α = 0.75 0.005) Canada Boys PA and BI Boys BI, r = -.07, (p < -0.08 0.005) Ransdell et al N = 14 girls Experimental PSPP PA recall ANOVA Although no significant increases in NR NR (2001) Age range 11-17 (12 Week pilot and 6 α = DNR questionnaire PA were reported, significant USA Month follow up) improvements in SC, PC, and STR were reported. BI did not improve

over time. Raudseep et al N = 253 Cross-sectional CY-PSPP PA self-report Bivariate All sub-domains of the CY-PSPP Girls Girls BA, 0.34; (2002) 119 girls and 134 boys α = 0.77 correlations (SC, STR, PC and PSW) were PA and BA r = Girls SC, 0.24; Age range 11-14 Multiple regression significantly (p < 0.05) associated 0.30, (p < 0.01) Girls STR, 0.19; Estonia with PA in both sexes. PA and SC r Girls PC, 0.23; 0.21, (p < 0.05) Girls GPSW, PA and STR r = 0.26 0.17, (p < 0.05) PA and PC r = Boys BA, 0.19; 0.20, (p < 0.05) Boys SC, 0.38; PA and GPSW r Boys STR, 0.42; = 0.23, (p < Boys PC, 0.35; 0.01) Boys GPSW, 0.34 Boys PA and BA r = 0.17, (p < 0.05) PA and SC r = 0.33, (p < 0.01) PA and STR r = 0.37, (p < 0.01) PA and PC r = 0.31, (p < 0.01) PA and GPSW r =0.30, (p < 0.01) Raustorp et al N = 77 Longitudinal CY-PSPP Pedometers Bivariate PA showed poor correlations to sub Girls Girls (2009) 41 girls and 36 boys (3 years) α = DNR correlations domains ad to PSW in both boys and PA and PC PC, 0.12; Mean ages girls. r = 0.10, (p = > PF, -0.11; T1= 12.7 0.05) BI, 0.08; T2 = 15.7 PA and PF GP, -0.10 T3 = 17.7 r = -0.09, (p = > Sweden 0.05) Boys PA and BI PC, 0.33; r = 0.07, (p = > PF, 0.16; 0.05) BI, 0.05; PA and GP GP, 0.31 r = -0.08, (p = > 0.05)

Boys PA and PC r = 0.28, (p = > 0.05)

PA and PF r = 0.13, (p = > 0.05) PA and BI r = 0.04, (p = > 0.05) PA and GP r = 0.26, (p = > 0.05)

Raustorp et al Year 2000 Cohort Longitudinal CY-PSPP Pedometers Logistic regression In girls, BA (r = 0.86, p < 0.001) NR NR (2006) N=501 (3 years) α = DNR showed the strongest correlation to 253 girls and 248 boys PSW. In boys, PC (r = 0.89, (p < Age range 7-14 0.001) showed the strongest Year 2003 Cohort correlation to PSW. N = 375 Age range 15-18 Sweden

Raustorp et al 2 Tests 2 Groups Longitudinal CY-PSPP Pedometers Bivariate PSW at T1 was a significant Girls Girls (2005) Group 1 N= 48 (3 years) α = DNR correlations for predictor of PA three years later in PA and SC r = BA, 0.22 27 girls and 21 boys cross-sectional data girls but not in boys. 0.19, (p < 0.05) Girls SC, 0.23; Age range 11-12 Logistic regression PA and BA r = Girls STR, 0.20; Group 2 N = 501 for longitudinal data 0.18, (p < 0.05) Girls GPSW, 253 girls and 248 boys PA and STR r = 0.16 Age range 10-14 0.17, (p < 0.05) Sweden PA and GPSW r Boys = 0.13, (p < BA, 0.36 0.05) SC, 0.42 Boys STR, 0.43; Boys Boys GPSW, PA and SC r = 0.32 0.35, (p < 0.05) PA and BA r = 0.30, (p < 0.05) PA and STR r = 0.36, (p < 0.05) PA and GPSW r = 0.27, (p < 0.05)

Rodriguez and N = 983 Cross-sectional from PSDQ Youth Risk Behaviour SEM PA had a significant effect on GPSC PA and GPSW GPSW, 0.18 McGovern 511 girls and 472 boys a longitudinal study α = 0.96 Surveillance for boys (β = 0.692, z = 4.51, p < r = 0.18, (p < (2005) Age - 17 0.001) and girls (β = 0.682, z = 5.00, 0.05) USA p < 0.001).

Scarpa and Nart N = 394 Cross-sectional PSDQ-S PA self-report Bivariate Positive associations between SC PA and END r = SC, 0.66; (2012) 221 girls and 173 boys α = DNR correlations scales and PA. 0.53, (p < 0.001) STR, 0.44 Age range 12-13 PA and FLX r = Mean age = 12.2 0.20, (p < 0.001) Italy PA and STR r = 0.37, (p < 0.001) PA and CRD r = 0.43, (p < 0.001) PA and SC r = 0.55, (p < 0.001) Schmalz et al N = 197 girls Longitudinal PSPP-C The Children’s Individual growth Participating in PA lead to positive NR NR (2007) Age range 9-13 (10 years) α = 0.78 for GSE Physical Activity scale model SE USA

Schneider N = 120 girls Experimental PSDQ PA recall HGLM Participation in vigorous exercise NR NR (2008) Age range 16-17 α = DNR questionnaire Bernoulli was dichotomized. The extent to USA distribution. which the intervention influenced vigorous PA was tested using the Hierarchical generalized Linear Model (HGLM) distribution and a non-linear analysis for binary outcomes using the Bernoulli distribution.

The interaction between group and change in vigorous PA was not significant for any of the PSDQ scales.

PA intervention among sedentary adolescent girls enhanced general physical self-concept for a subset of intervention participants who manifested positive change in fitness. Schneider and N = 192 Experimental PSDQ PA recall Multilevel analyses Greater in PA lead to increases in NR NR Kwan (2013) 87 and 105 boys (9 Months) α = DNR GSC Age range 14-16 Mean age = 14.78, SD =0.45 USA

Shriver et al N = 214 Cross-sectional Body Esteem Scale SAPAC ANOVA No significant correlations were NR NR (2013) 99 girls and 115 boys α = 0.83 found between physical activity Age range 8-10 as a continuous variable (minutes USA per day) and body esteem scores (global, BE-Weight, BE- Appearance Slater et al N = 714 Cross-sectional Appearance Anxiety PA recall Bivariate Participation in PA that had a strong Girls aesthetic Girls PA (2011) 332 girls and 382 boys Scale correlations focus on appearance is related to PA and aesthetic, -.01 Age range 12-16 Girls, α = 0.91 higher levels of body image appearance Girls PA non Mean age = 14.47, SD =0.62 Boys, α = 0.88 concerns. anxiety aesthetic, -.04 Australia r = -0.01, (p > Boys PA 0.05) aesthetic, 0.10 Girls non- Boys PA non

aesthetic PA and aesthetic, 0.-.04 appearance anxiety r = -0.04, (p > 0.05) Boys aesthetic PA and appearance anxiety r = 0.09, (p > 0.05) Boys non- aesthetic PA and appearance anxiety r = -0.04, (p > 0.05) Smart et al N = 222 girls Cross-sectional CY-PSPP PAQ-A SEM PA was positively associated with (SEM) PA and SEM (2012) Age range 10-14 α = 0.91 Bivariate all dimensions of the CY-PSPP and PSW SC, 0.53; Mean age = 12.7, SD = 0.8 correlations all findings were statistically r = 0.50, (p < England significant. 0.05) Correlation BI, 0.35; PA and SC STR, 0.42; r = 0.51, (p < PSW, 0.45 0.001) PA and PC r = 0.52, (p < 0.001) PA and BI r = 0.33, (p < 0.001) PA and STR r = 0.40, (p < 0.001) PA and PSW r = 0.43, (p < 0.001) Sollerhead et al N=206 Cross-sectional Perception and PA recall Logistic regression Physically active children had more NR NR (2008) 92 girls and 114 boys Confidence questionnaire positive self-perceptions and SC. Age range 8-12 Questionnaire Sweden α = DNR

Spink et al N = 153 Cross-sectional Potential correlates Modified Activity Bivariate Frequency of PA was strongly NR NR (2006) 67 girls and 86 boys questionnaire Questionnaire correlations influenced by SC Age range 12-17 for Adolescents Mean age = 15.5, SD = 1.68 Canada

Standage et al N = 494 Cross-sectional PSDQ Pedometers Bivariate SC and BA approached PA and SC SC, 0.12; (2012) 291 girls, 201 boys and 2 did SC α = 0.89 correlations SEM significance with regard to PA r = 0.11, (p < BA, 0.04 not specify sex. BA α = 0.88 relating PE involvement 0.05) Mean age = 12.58, SD =0.74 Perceived PA and BA England Competence subscale r = 0.04, (p > of the Intrinsic 0.05) Motivation Inventory Stein et al N = 8670 Cross-sectional PSPP-C PA recall Linear regression Linear regression was used to ∆ PA and SC in Girls SC, 0.08 (2007) 5260 girls and 3410 boys α = DNR questionnaire examine the effect of continuous girls Age range 9-14 change in PA on continuous change r = 0.07, (p < Boys SC, 0.06 USA in each area of self-perception over 0.01) the study period, adjusting for age, ∆ PA and SC in BMI, baseline activity level, boys baseline self-perception scores, past- r = 0.05, (p < month smoking, and exercising for 0.01) weight control.

For both girls and boys, linear regression models showed that increase in physical activity was positively associated with change in social and athletic self-perception scores (p < 0.01) but was not associated with a change in either scholastic or global self-perception. Sullivan (2002) N = 1602 Cross-sectional PCSC PA recall Multiple regression A positive association was evident NR NR 810 girls and 792 boys α = DNR questionnaire between PA and SC for both girls Age range 11-12 and boys. Ireland

Taylor et al N = 178 Longitudinal Psychological Need PAQ-C Intraclass People with higher SC report greater NR NR

(2010) 55 girls and (3 years) Satisfaction Correlation PA. 123 boys questionnaire Coefficients Age range 11-16 Mean age = 13.8, SD = 1.29 UK Trautwein et al Year 2001 Cohort Longitudinal SPPC PA recall Multiple regression T1 PA was significantly (p < 0.05) NR NR (2008) N = 1185 (15 months) questionnaire associated with T2 PSC after 675 girls and 510 boys adjusting for sex, BMI, PE grade, T1 Mean age = 9.67 PSC and physical ability. 2002 N = 1095 T1 PSC was significantly (p < 0.05) Germany associated with T2 PA after adjusting for sexes, BMI, peer grade, T1 PA and physical ability.

Trew et al N = 602 Cross-sectional SPPA PA recall Cluster analysis Boys spent more time participating NR NR (1999) Age range 11-18 questionnaire in sport than girls and boys reported Ireland higher perceptions of ‘athletic self- competence’ and ‘global self-worth’ than girls. Vierling et al N = 239 Cross-sectional Intrinsic Motivation Pedometers Bivariate PA was positivity and significantly PA and SC SC, 0.20 (2007) 119 girls and 120 boys Inventory correlations associated with SC. r = 0.18, (p < Age range = 9.81-14.41 0.01) Mean age = 12.11, SD = 1.21 Self-regulation (SEM) USA Questionnaire SEM included correlations Children’s Attraction among to Physical Activity autonomy, scale competence and α relatedness. SC α Viira (2011) N = 203 Longitudinal CY-PSPP PA self-report Longitudinal Important predictors PA for boys NR NR 105 girls and 98 boys (12 months) regression analyses were PC and PSW. For girls BI and Mean age = 15.2, SD = 0.57 PC were important predictors of PA. Estonia

CY-PSPP sub-domains are important determinants of PA among early adolescent’s boy’s perceptions of COND, SPORT, BODY, STR, PSW and GSW were higher and they participated more in MVPA compared to girls. Wang and N = 2510 Cross-sectional Task and ego PA recall Bivariate Positive correlation between PA and PA and SC r = SC, 0.63; Biddle (2001) 1332 girls and 1178 boys orientation in sport questionnaire correlations PC as well as PA and GPSW. 0.56, (p < 0.01) GPSW, 0.46 Age range 11-15 questionnaire PA and GPSW r Mean age = 12.95, SD = 0.90 Conceptions of the = 0.43, (p < England nature of athletic 0.01) ability questionnaire PSPP-C

Wang et al N = 824 Cross-sectional PSPP-C PA self-report Bivariate Positive correlation between PA and PA and SC SC, 0.43 (2002) 391 girls and 427 boys correlations SC r = 0.38, (p < 6 did not specify sex 0.001) Age range 11-14 Mean age = 12.71, SD =0.87 England Wang et al N = 984 Cross-sectional PSPP-C PA self-report Bivariate Positive correlation between PA and PA and SC SC, 0.33 (2010) 642 girls and 342 boys correlations SC r = 0.29, (p < Age range 11-19 0.01) Mean age = 14.12, SD = 3.11 China

Welk and N = 25 Experimental CY-PSPP Accelerometers Bivariate Correlations between PA and SC PA and SC Week 1 SC, Schaben (2004) 11 girls and 14 boys (3 weeks) α = 0.84 correlations were high for all three weeks (mean week 1 0.63; Age range 8-12 r = 0.72). r = 0.58, (p < Week 2 SC, USA 0.05) 0.75; PA and SC Week 3 SC, week 2 0.62; r = 0.69, (p < Week 4 SC, 0.79 0.05) PA and SC week 3

r = 0.57, (p < 0.05) PA and SC week 4 r = 0.72, (p < 0.05)

Welk et al N = 994 Cross-sectional PSPP PAQ-C Bivariate Strong and significant correlations Girls PA and SC Girls SC, 0.45; (2003) 489 girls and 505 boys Girls SC α = 0.72 PA Recall correlations between PA and SC was present r = 0.38, (p < Boys SC, 0.54; Mean age 9.95 questionnaire 0.05) Total SC, 0.51 USA Boys SC α = 0.70 Boys PA and SC Total SC α r = 0.46, (p < 0.05) Total PA and SC r = 0.43, (p < 0.05) Zhang et al N = 286 Experimental PE modified health PA questionnaire for Bivariate Positive association between PA and PA and SC r = SC, 0.53 (2011) 143 girls and 143 boys care climate older children correlations SC 0.44, (p < 0.01) Mean age = 13.4, SD = 1.0 questionnaire USA Perceived needs satisfaction α = DNR 60 61 62 BA = Body Attractiveness. BD = Body Dissatisfaction. BI = Body Image. CRD = Co-ordination. CY-PSPP = Children and Youth Physical Self-Perception Profile. DNR = Did not report. EA = External Attribution. 63 END = Endurance. FLX = Flexibility. GPSC = General Physical Self Concept. GPSE = Global Physical Self Esteem. GSE = Global Self Esteem. GPSW = Global Physical Self Worth. LTEQ = Leisure Time Physical 64 Activity Questionnaire. N = Number of participants. PA = Physical Activity. PAAQ = Physical Activity Assessment Questionnaire. PAQ-A = Physical Activity Questionnaire for Adolescents. PAQ-C = Physical 65 Activity Questionnaire for Children. PC = Physical Conditioning. PCSC = Perceived Competence Scale for Children. PE = Physical Education. PHCSCS = Piers-Harris Children’s Self-Concept Scale. PSCS = Physical 66 Self Concept Subscale. PSDQ-S = Physical Self-Description Questionnaire-Short. PSPP = Physical Self Perception Profile. PSPP-A = Self-Perception Profile for Adolescents. PSPP-C = Physical Self-Perception 67 Profile for Children. PSWS = Physical Self Worth Scale. SAPAC = Self-Administered Physical Activity Checklist. SC = Sports Competence. SDQ = Self-Description Questionnaire. SEM = Structural Equation Model. 68 SP = Self-Perception. SPA = Social Physique Anxiety Scale. SPAA = Harter’s Self-Perception Profile for Adolescents. SPAS = Social Physique Anxiety Scale. SPPA = Self-Perception Profile for Adolescents. SPPC = 69 Self-Perception Profile for Children. STR = Physical Strength. SS = Sport Skill.  = Standardized beta co-efficient. 70 a = Cronbach’s alpha α .

37 71 Electronic Supplementary Table S2: PRISMA checklist table Reported Section/topic # Checklist item on page # TITLE Title 1 Identify the report as a systematic review, meta-analysis, or both. 1 ABSTRACT Structured summary 2 Provide a structured summary including, as applicable: background; objectives; data sources; 3,7-9 study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number. INTRODUCTION Rationale 3 Describe the rationale for the review in the context of what is already known. 5-7 Objectives 4 Provide an explicit statement of questions being addressed with reference to participants, N/A interventions, comparisons, outcomes, and study design (PICOS). METHODS Protocol and 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if N/A registration available, provide registration information including registration number. Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., 7-8 years considered, language, publication status) used as criteria for eligibility, giving rationale. Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study 7-10 authors to identify additional studies) in the search and date last searched. Search 8 Present full electronic search strategy for at least one database, including any limits used, such 9 that it could be repeated. Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, 7-10 and, if applicable, included in the meta-analysis). Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) 9-10 and any processes for obtaining and confirming data from investigators.

38 1 Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any N/A assumptions and simplifications made. Risk of bias in 12 Describe methods used for assessing risk of bias of individual studies (including specification of 9,13- individual studies whether this was done at the study or outcome level), and how this information is to be used in 14,20,76- any data synthesis. 77 Summary measures 13 State the principal summary measures (e.g., risk ratio, difference in means). 7,10 Synthesis of results 14 Describe the methods of handling data and combining results of studies, if done, including 10-12 measures of consistency (e.g., I2) for each meta-analysis. 72 Reported Section/topic # Checklist item on page # Risk of bias across 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication 12-14,76- studies bias, selective reporting within studies). 77 Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), 12 if done, indicating which were pre-specified. RESULTS Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with 40 reasons for exclusions at each stage, ideally with a flow diagram. Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, 45-75 follow-up period) and provide the citations. Risk of bias within 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see 19-21, 76- studies item 12). 77 Results of individual 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data 41-44 studies for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot. Synthesis of results 21 Present results of each meta-analysis done, including confidence intervals and measures of 15-18,41- consistency. 44

39 2 Risk of bias across 22 Present results of any assessment of risk of bias across studies (see Item 15). 20,76-77 studies Additional analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta- 2,15-18 regression [see Item 16]).

DISCUSSION

Summary of evidence 24 Summarize the main findings including the strength of evidence for each main outcome; consider 2,30-31 their relevance to key groups (e.g., healthcare providers, users, and policy makers). Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., 20-21,31 incomplete retrieval of identified research, reporting bias). Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications 3,30-31 for future research. FUNDING Funding 27 Describe sources of funding for the systematic review and other support (e.g., supply of data); 32 role of funders for the systematic review. 73

40 3 74 Electronic Supplementary Table S3 – Quality assessment of studies meeting inclusion 75 criteria 76 Study 1 2 3 4 5 Abarca-SOS et al (2013)      Annesi (2006)      Annesi et al (2008)      Baker & Davison (2011)      Barnett et al (2011)      Barnett et al (2008)      Bevans et al (2010)      Biddle and Goudas (1996)      Biddle & Wang (2003)      Chen et al (2010)      Cheng et al (2003)      Craft et al (2003)      Crocker et al (2003)      Crocker et al (2006)      Crocker et al (2000)      Cumming et al (2011)      Dishman et al (2006)      Douthitt (1994)      Duncan et al (2004)      Duncan et al (2006)      Dunton et al (2003)      Dunton et al (2006)      Eccles & Harold (1991)      Gillison et al (2011)      Goldfield et al (2011)      Guinn et al (1997)      Haugen et al (2013)      Ingledew & Sullivan (2002)      Jaakkola et al (2013)      Jackson et al (2013)      Kalaja et al (2010)      Knowles et al (2009)      Lau et al (2004)      Lau et al (2006)      Loucaides et al (2004)      Lubans et al (2011)      Luszcynsk & Abraham (2012)      Malete et al (2008)      Markland and Ingledew (2007)      Marsh (1996)      Martin et al (2006)      Moreno-Murcia et al (2011)      Morgan et al (2008)      Niven et al (2009)      Niven et al (2007)     

41 4 Study 1 2 3 4 5 Papaioannou et al (2006)      Paxton et al (2004)      Plotnikoff et al (2007)      Raudseep et al (2002)      Raustorp et al (2009)      Raustorp et al (2005)      Rodriguez & McGovern (2005)      Scarpa & Nart (2012)      Slater et al (2006)      Smart et al (2012)      Standage et al (2012)      Stein et al (2007)      Vierling et al (2007)      Wang & Biddle (2001)      Wang et al (2002)      Wang et al (2010)      Welk and Schaben (2004)      Welk et al (2003)      Zhang et al (2011)      77  - Explicitly described and present;  - absent or unclear or inadequately described.