at SciVerse ScienceDirect

Psychology of Sport and Exercise 14 (2013) 447e454

Contents lists available

Psychology of Sport and Exercise

journal homepage: www.elsevier .com/locate/psychsport

Biological maturation and physical activity in adolescent British females: The rolesof physical self-concept and perceived parental support

Laura Jackson a, Sean P. Cumming a,*, Clemens Drenowatz b, Martyn Standage a, Lauren B. Sherar c,Robert M. Malina d,e

aDepartment for Health, University of Bath, Bath, UKbDivision of Sports and Rehabilitation, Ulm University Medical Centre, Germanyc School of Sport, Exercise and Health Sciences, University of Loughborough, UKdDepartment of Kinesiology and Health Education, University of Texas at Austin, TX, USAeDepartment of Kinesiology, Tarleton State University, TX, USA

a r t i c l e i n f o

Article history:Received 23 April 2012Received in revised form20 November 2012Accepted 2 January 2013Available online 1 February 2013

Keywords:PubertyGirlsExerciseMaturity timingParentsAttractiveness

* Corresponding author.E-mail address: s.cumming@bath.ac.uk (S.P. Cumm

1469-0292/$ e see front matter � 2013 Published byhttp://dx.doi.org/10.1016/j.psychsport.2013.01.001

a b s t r a c t

Objectives: This study examined the mediating and moderating effects of physical self-concept andparental support, respectively, on relations between maturation and physical activity (PA) in Britishadolescent females.Design: Cross-sectional field based study.Methods: Biological maturity status, physical self-concept, perceived parental support for PA, and self-reported PA were assessed in 244 female British year 7e9 pupils (M age ¼ 12.8 years, SD ¼ .9).Results: Structural equation modelling, employing maximum likelihood estimation and boot-strappingprocedures supported the contention that physical self-concept mediated an inverse relation betweenmaturation and PA. A regression model examining the main and interactive effects of maturation andparental support on PA provided evidence of a main effect for parental support, but no interactive effect.Conclusions: The results suggest that physical self-concept partially mediates an inverse relation betweenmaturity and physical activity in adolescent females. Accordingly, how adolescent females interpret orperceive the changes associated with maturation may be more important than maturation itself.Encouraging adolescent females to view puberty as a natural and attractive aspect of the process of‘growing up’ may help mitigate any negative health consequences associated with early maturation.Parental support for PA does not appear to moderate relations between maturation and PA.

� 2013 Published by Elsevier Ltd.

Regular involvement in physical activity (PA) is a prerequisite forhealthy physical and psychological development (Malina, 2008).Individuals who regularly engage in moderate-to-vigorous in-tensity bouts of PA demonstrate improved functional capacity andare at lower risk for numerous degenerative diseases and psycho-logical disorders (e.g., hypertension, type II diabetes, coronary heartdisease, depression & anxiety). To attain the health benefits of PA,individuals should be encouraged to be, and remain, active from anearly age. Children have, however, become less active in many PAdomains, including physical education, active transport, andleisure-time activities (Dollman, Norton, & Norton, 2005).

A true understanding of PA in youth resides in the independentand interactive effects of biological, psychosocial, and environmental

ing).

Elsevier Ltd.

factors (Eisenmann & Wickel, 2009). Whereas the contributions ofpsychological and environmental factors are well documented, theinfluence of biological factors has received less attention. A biologicalfactor relevant to PA in youth is maturation. Biological maturationentails progression towards adulthood and can be considered interms of tempo and/or timing (Malina, Bouchard, & Bar-Or, 2004).Tempo refers to the rate at which maturation progresses, whereastiming refers to the time that maturity-related events occur relativeto same age peers (e.g., age-at-menarche or peak height velocity).Children of the same chronological age can vary considerably inmaturity, with some maturing earlier or later than others.

The timing of biological maturation may contribute to PA inadolescence, particularly in females where marked declines in PAand increases in sedentary behaviours coincide with the onset ofpuberty (Sallis, 2000). An inverse relation betweenmaturity timingand PA is predicted in adolescent females. The physical and func-tional characteristics associated with advanced maturation in girls

L. Jackson et al. / Psychology of Sport and Exercise 14 (2013) 447e454448

(i.e., greater pubertal gains in height, weight, weight-for-height,and absolute and proportional fat-mass) are less conducive tosuccessful engagement in PA (Sherar, Cumming, Eisenmann,Baxter-Jones, & Malina, 2010), especially in activities that empha-sise weight bearing, endurance, and/or aesthetic qualities (e.g.,diving, gymnastics, figure skating). Early maturing girls also so-cialize with older peers groups (Caspi & Moffitt, 1991), who by theirnature, are less active. Conversely, late maturing girls socialize withyounger and more active peers, and possess physiques that aremore suited to successful engagement in PA and more consistentwith western ideals pertaining to attractiveness (i.e., more linearphysique, less absolute and proportional fat-mass).

Studies of maturity timing and PA in adolescent females haveproduced equivocal results (Sherar et al., 2010). Whereas somestudies have shown late maturing females to be more active thanearly maturing females (Davison, Werder, Trost, Baker, & Birch,2007; Riddoch et al., 2007), an equivalent number of studies haveobserved no relation between maturity timing and PA (Niven,Fawkner, Knowles, & Stephenson, 2007; Wickel & Eisenmann,2007). That said, only one study, found early maturing girls to bemore active (van Jaarsveld, Fidler, Simon, & Wardle, 2007). Theinconsistent nature of these findings has been attributed to thefailure to consider moderating and mediating factors (Sherar et al.,2010).

To better understand relations between maturation and PA inadolescence, Cumming, Sherar, et al. (2011) proposed a BioculturalModel of Maturity-associated Variance in Adolescent PA. In linewith existing models of pubertal and adolescent development(Holmbeck, 2002; Petersen & Taylor, 1980), the Biocultural Modelholds that maturation exerts both direct and indirect (i.e., mediated& moderated) effects on PA. Mediating variables include beliefs,fantasies and/or attitudes towards maturation, the body, or PA;whereas moderating factors include peer acceptance, parentalsupport, motivational climate, and cultural values and standardspertaining to the body, maturation and PA.

A central tenet of the Biocultural Model is that the physical andfunctional changes associated with puberty hold significant stim-ulus value, influencing adolescents’ self perceptions and the per-ceptions and reactions of others. From this perspective, it is not somuch the physical and functional changes associated with pubertythat influence adolescent engagement in PA, but the manner inwhich the changes are interpreted and the meaning and value thatis ascribed to them (Cumming, Sherar, et al., 2011; Petersen &Taylor, 1980). In line with this reasoning, individual differences inthe activity of early maturing girls are attributed to differences inthe interpretation and social management of puberty. That is, earlymaturing girls who perceive pubertal change as normal andattractive consequence of the maturation process (and not asa barrier to activity) and/or experience an environment that issupportive and accepting of change (i.e., autonomy supportive,mastery involving, or body image friendly) are considered to bemore likely to remain active through adolescence.

Evidence suggests that perceptions of the physical-self mediatean inverse relation between maturation and PA in adolescent girls.In separate samples of adolescent British females, Cumming,Standage, et al. (2011) and Hunter Smart et al. (2012) tested a se-ries of mediated effects models explaining maturity associatedvariance in PA. Structural equation modelling, employing boot-strapping procedures, provided strong support for the models andevidence of mediation effects. In each study, advanced maturationpredicted lower physical self-concept, which, in turn, predicted lessinvolvement in PA. Maturation and physical self-concept weremostclosely associated with perceptions of attractiveness and, to a lesserextent, sports competence. Early maturing girls reported lowerattractiveness and sports competence, which, in turn, predicted

a more negative self-concept and less involvement in PA. The issuesof attractiveness and self-presentation are considered to play cen-tral roles in the psycho-behavioural adjustment of adolescent fe-males and, to a lesser extent, males (Summers-Efler, 2004). Theneed to feel and demonstrate competence has also been identifiedas a key predictor of motivated behaviour and well-being ina number of contemporary theories of motivation, including Har-ter’s Competence Motivation Theory (Harter, 1981) and, morerecently, Self-determination Theory (Ryan & Deci, 2000).

Parental support for PA may serve as a moderator of relationsbetween maturation and PA in adolescent females. It consists ofa variety of behaviours including encouragement, role modelling,logistical support, and active engagement, expressing interest, and/or watching/spectating, and has been shown to predict adolescentinvolvement in PA. Mrug et al. (2008) observed that early maturinggirls who experienced positive parenting were less susceptible toengagement in delinquent behaviours. In a similar fashion, parentalsupport for PA may mitigate any negative effects of early matura-tion on PA.

A limited number of studies have examined the impact ofparental support on relations between maturation and PA. Themain and interactive effects of parental support and maturationon moderate-to-vigorous intensity PA (MVPA) were examined in801 participants in the NICHD Study of Early Child Care and YouthDevelopment (Bradley, McRitchie, Houts, Nader, & O’Brien, 2011).Though pubertal timing was unrelated to MVPA in both malesand females, parental support moderated the effect of pubertaltiming on MVPA in males, but not females. Specifically, highparental monitoring predicted lower activity levels for latematuring, but greater activity levels for early maturing males. Ina study of early breast development in adolescent girls, Summers-Efler (2004) noted that high levels of parental acceptance andsupport predicted continued involvement in PA through ado-lescence. This observation, amongst others, led Summers-Efler tocontend that parental support played a central role in the pre-vention and development of defensive or maladaptive copingstrategies and the internalization of negative messages in earlymaturing girls.

In light of Sherar et al.’s (2010) recommendation to considervariables that might mediate or moderate relations betweenmaturation and PA, the purpose of this study was to examine theextent to which physical self-concept and perceptions of parentalsupport for PA mediated and moderated, respectively, relationsbetween maturation and PA in adolescent females. Consistent withprevious research, it was predicted that physical self-conceptwould mediate an indirect inverse relation between maturationand PA. Further, it was predicted that perceptions of parentalsupport for PA would moderate relations between maturation andPA. That is, high levels of perceived parental support would miti-gate any negative effects associated with advanced maturationon PA.

Method

Participants

Participants were 244 female Year 7 through 9 pupils froma single-sex Catholic Comprehensive State School in England (Mage ¼ 12.8, SD ¼ .9 years; range ¼ 11e14 years). The study’smethods were approved by the Department for Health’s researchethics committee. Written consent was obtained from the HeadTeacher, who acted in loco parentis. Parents were informed of theresearch by post and asked to provide passive consent (i.e., contactthe school/researchers if they did not wish their child to take part).Verbal consent was obtained from pupils.

L. Jackson et al. / Psychology of Sport and Exercise 14 (2013) 447e454 449

Measures

Estimated maturity statusPercentage of predicted mature (adult) height attained at

measurement served as a non-invasive estimate of biologicalmaturity status. This approach assumes that among children of thesame chronological age, the child who is closer to his/her predictedmature height is more advanced in biological maturity. Conversely,the individual who is further removed from his/her predicted adultheight than expected for age is considered less biologically mature(Malina et al., 2004). For example, the mean percentage of matureheight attained in girls of the Berkeley Guidance Study at the age of12 years is 93% (Bayer & Bayley, 1959). A girl who has attained 98%of her predicted adult height at 12 years would be consideredbiologically more mature than a girl of the same chronological agewho has only attained 86% of her predicted adult height.

The KhamiseRoche (KR) method (Khamis & Roche, 1994) wasemployed to predict the mature height from current age, heightand weight of the participant and midparent height (i.e., meanheight of biological parents). The median error bound (medianabsolute deviation) between actual and predicted mature height at18 years of age is 2.2 cm in males and 1.7 cm in females (Khamis &Roche, 1994).

Estimated biological maturity status was expressed as a z-score,using the percentage of predicted adult height attained at the timeof measurement, and half-year age- and sex-specific means andstandard deviations from the Berkeley Guidance Study (Bayer &Bayley, 1959). The KhamiseRoche method for estimating bio-logical maturation has been used in a number of studies with USand British youth, demonstrating concurrent validity (Cumming,Battista, Standage, Ewing, & Malina, 2006; Malina, Morano,Barron, Miller, & Cumming, 2005; Sweet, Dompier, Stoneberg, &Ragan, 2002). It has also been validated against established in-dicators of maturity (skeletal age) in youth American footballplayers (Malina, Dompier, Powell, Barron, & Moore, 2007). Fora more comprehensive account of the method used to estimatematuration please see Cumming, Standage, et al. (2011).

Physical activityThe Physical Activity Questionnaire for Adolescents (PAQ-A)

(Kowalski, Crocker, & Donen, 2004) was used to assesses involve-ment in PA. The PAQ-A is an eight item instrument that requiresparticipants to indicate how frequently they engaged in variousactivities over a seven day recall period. The PAQ-A has demon-strated adequate levels of validity and reliability (Kowalski, Crocker,& Kowalski, 1997). In the current sample the PAQ-A demonstratedan acceptable level of internal consistency (Cronbach’s alpha¼ .75).

Physical self-conceptThe Children and Youths’ Physical Self-Perception Profile (CY-

PSPP) (Whitehead, 1995) was used to assess the physical self-concept. This CY-PSPP assesses 6 dimensions of the self (i.e., sportcompetence, physical condition, body attractiveness, strength,physical self-worth, and general self-worth) and entails partici-pants responding to 36 items structured in an alternative responseformat Theoretically, the physical self-worth dimension is deemeda higher-order construct, with sport competence, physical condi-tion, body attractiveness, and strength serving as lower-orderconstructs that contribute to physical self-worth. All dimensionsof the CY-PSPP except general self-worth were used in the currentstudy. The CY-PSPP has demonstrated acceptable levels of constructvalidity and reliability in previous research (Eklund, Whitehead, &Welk, 1997). Further, each dimensions of the CY-PSPP demon-strated adequate levels of internal consistency in the currentsample (Cronbach’s alphas ranged from .81 to .87).

Parental support for physical activityA PA specific version of the Parental Support Scale (Trost et al.,

2003) was used to measure the participants’ perceptions ofparental support for PA. The five-item scale was completed by theparticipants and assessed parental (a) encouragement, (b) praise,(c) transportation of the child, (d) PA participation with the child,and (e) watching the child participation in PA or sports. Fre-quency of parent support was assessed on a 5-point Likert scaleranging from never (0) to everyday (4). Total scores were summedwith high scores reflecting greater perceptions of parental sup-port for PA. Trost et al. (2003) found the scale to demonstrateadequate levels of test-retest reliability (r ¼ .81) and internalconsistency (Cronbach’s alpha ¼ .78). The scale also demonstratedadequate internal consistency (Cronbach’s alpha ¼ .84) in thecurrent study.

Field protocol

As part of a Physical Education class, participants completedthe self-report questionnaires. Height and weight were measuredusing standardized procedures (Malina, 1995). Chronological agein decimals was calculated as the difference between date of birthand date of measurement. The biological height of the parentswas collected through a self-report form that was completed athome and returned by the participants. As adults generallyoverestimate height, the self-reported height of each parent wasadjusted for over estimation (Epstein, Valoski, Kalarchian, &McCurley, 1995).

Statistical analyses

Descriptive statistics by age group were calculated for body size,percentage of predicted mature height, maturity status, physicalself-concept, PA, and perceived parental support for PA. Pearsonproduct moment correlations (one-tailed) were calculated toexamine relations among the variables of interest. Structuralequation modelling, utilizing maximum likelihood estimation andbootstrapping procedures, was used to test the hypothesizedmodelconcerning direct and indirect relations amongmaturational status,physical self-concept, and PA. To determine the adequacy of modelfit, a 2-index presentation strategy advanced by Hu and Bentler(1999) was employed. This strategy uses the Standardized RootMean Square Residual (SRMR) and an incremental or absolute in-dex of fit (e.g., Comparative Fit Index e CFI). SRMR values close to.08 (or lower) are indicative of awell-specifiedmodel (Hu & Bentler,1999) whereas CFI values of over .90 and .95 reflect acceptable andexcellent fit between the model and data, respectively (Hu &Bentler, 1999). In accordance with recent recommendations(MacKinnon, Lockwood, & Williams, 2004), mediated effects wereexplored by examining the 90% upper and lower limits ofbootstrap-generated bias-corrected confidence intervals (BBC CI) ofindirect effects.

The main and interactive effects of maturation and parentalsupport for PA on PAwere examined via two-step forced-entry, fullhierarchical regression model (Type III Sum of Squares was used totest the hypotheses). The participants age was entered as a predic-tor in step one of the analysis. Maturity status and perceivedparental support were entered as predictor in step two (Malinaet al., 2004). Following Aiken and West (1991), we centred thematurity status and perceived parental support values by sub-tracting the sample mean for the variable from each individual’svalue. The centred values were used to generate the interactionterm for the purpose of reducing collinearity between main andinteraction effects. The interaction term was entered in step threeof the model.

Table 2Pearson product moment correlations (one-tailed) between measures of estimatedbiological maturity status, physical self-concept, physical activity, and parentalsupport for physical activity (PA) in adolescent British females.

1 2 3 4 5 6 7

1. Biological maturitystatus

2. Sport competence �.073. Physical condition �.26*** .67***4. Body attractiveness �.35*** .43*** .52***5. Physical strength .14* .45*** .35*** .20***6. Physical self-worth �.28*** .63*** .65*** .76*** .51***7. Physical activity �.20** .48*** .50*** .20** .40*** .37***8. Parental support

for PA�.22*** .39*** .36*** .19** .26*** .35*** .54***

*p < .05; **p < .01; ***p < .001.

L. Jackson et al. / Psychology of Sport and Exercise 14 (2013) 447e454450

Results

Descriptive statistics

Descriptive statistics for chronological age, body size, estimatedmaturity status, physical self-concept, PA and perceived parentalsupport are presented by age group in Table 1. Mean values formaturity status z-scores approximated, or fell just below, 0 in the11,12, and 13 year old age groups andwere lowest in the 14 year oldage group. Compared to UK reference values (Child GrowthFoundation, 2005) mean height values approximated at the 75thcentile at age 11, 12, and 13 and fell between the 50th and 75thcentiles at 14. Mean values for weight approximated the 75thcentile at age 11, 12, and 14 and the 91st centile at age 13. Meanvalues for BMI fell between the 50th and 75th at age 11, 12, and 14,yet fell between the 75th and 91st centiles at age 13.

Correlations

Relations between biological maturity status, physical self-concept, and PA are presented in Table 2. As predicted, estimatedbiological maturity status was negatively associated with, physicalcondition, body attractiveness, physical self-worth, PA and pos-itively associated with physical strength. The magnitude of theinverse relation betweenmaturation and PAwas small tomoderate.The magnitude of the statistically significant correlations betweenmaturation and physical self-concept varied from small (physicalstrength, r ¼ �.14) to moderate (body attractiveness, r ¼ �.35).Though negative, the relation between maturation status and sportcompetence was non-significant. PA was positively associated withall dimensions of the CY-PSPP.

The mediated effects model

Consistent with the methods employed by Cumming, Standage,et al. (2011) and Hunter Smart et al. (2012) a parcelling strategy(Little, Cunningham, Shahar, & Widaman, 2002) was employed tolimit the number of estimated parameters. The six items on eachsubscale of the CY-PSPP were parcelled into three indicators ofa latent variable representing the subscale. The three items withthe highest item-to-construct loadings anchored each of the threeindicators. The three items with the next highest loadings weresubsequently added to the anchors in reverse order (e.g., 1st & 6th,2nd & 5th, & 3rd & 4th, highest loading items parcelled together),

Table 1Descriptive statistics for chronological age, estimated biological maturity status,physical self-perceptions, and physical activity of adolescent British females by agegroup.

11 yearsn ¼ 54

12 yearsn ¼ 87

13 yearsn ¼ 74

14 yearsn ¼ 29

Chronological age 11.71 (.2) 12.51 (.3) 13.52 (.3) 14.21 (.1)Maturity status

z-score.25 (.1.33) .15 (1.18) .37 (1.20) �.11 (.83)

Height (cm) 152.0 (9.1) 156.9 (7.9) 161.1 (6.7) 161.9 (6.0)Weight (kg) 46.4 (14.0) 50.0 (12.8) 59.8 (14.6) 56.0 (10.5)BMI (kg/m2) 19.7 (4.1) 20.2 (4.1) 22.9 (5.0) 21.5 (4.4)Sport competence 2.69 (.46) 2.62 (.47) 2.56 (.43) 2.48 (.53)Physical condition 2.71 (.55) 2.75 (.68) 2.59 (.53) 2.76 (.72)Body attractiveness 2.54 (.62) 2.53 (.70) 2.21 (.57) 2.36 (.71)Physical strength 2.61 (.60) 2.57 (.52) 2.42 (.61) 2.26 (.52)Physical self-worth 2.71 (.43) 2.66 (.64) 2.38 (.58) 2.52 (.70)Physical activity 2.55 (.51) 2.44 (.56) 2.25 (.51) 2.71 (.66)Parental support 9.41 (4.54) 7.55 (4.70) 6.64 (4.5) 7.34 (5.0)Percentage of

predicted adultstature attained

92.5 (4.4) 95.0 (3.3) 97.8 (2.1) 98.3 (1.0)

limiting experimenter bias, and increasing the probability that theparcels were balanced in terms of difficulty and discrimination(Little et al., 2002). The total score of the paired items was calcu-lated to create each indicator of the latent variable (i.e., parcelleditem). Employing the same strategy, the eight items representingthe PAQ-A were parcelled into four indicators of a latent variablereflecting PA. For an explanation of the parcelling process and itsadvantages see the review by Little et al. (2002).

As moderate-to-strong correlations were observed among thefour sub-dimensions of the CY-PSPP, covariance paths between thedisturbance terms of the latent factors were added to the hypoth-esized model. These paths specify the interrelations between thevarious sub-dimensions and the shared variance not accounted forby the predictor variable.

Employing AMOS 18.0 software, structural equation modelling(SEM) usingmaximum likelihood estimationwas used to test the fitof the hypothesized model. Inspection of the Mardia’s Coefficientvalue (73.43, p < .001) revealed the data to depart from multi-variate normality. Thus, the SEM analysis was conducted using thebootstrapping procedure with 5000 bootstrap replication samplesto provide a more accurate assessment of the parameter estimates’stability (Byrne, 2001).

The model fit indices suggested an acceptable fit between theproposed model and the data (SRMR ¼ .08; CFI ¼ .94). The stan-dardized Beta coefficients, standard errors, and squared multiplecorrelations (SMC) associated with the model are presented inFig. 1. The magnitude of the standardized path coefficients rangedfrom small (b < .10) to large (b > .50) (Kline, 2005; Suhr, 2006). Ashypothesized, biological maturity status negatively predicted per-ceptions physical condition (b ¼ �.29, p < .001) and body attrac-tiveness (b ¼ �.38, p < .001). As expected, maturity statuspositively predicted perceived strength (b ¼ .15, p < .05). Maturitystatus was, however, unrelated to perceptions of sport competence.Perceptions of body attractiveness served as the only positive andsignificant predictor of physical self-worth (b ¼ .57, p < .001).Physical self-worth, in turn, positively predicted involvement in PA(b ¼ .45, p < .001). Finally, maturity status was found to be indi-rectly related to physical self-worth (b¼�.28 (90% CI�.10), p< .01)and PA (b ¼ �.13 (90% CI �.06), p < .01).

A version of Baron and Kenny’s (1986) method for testingmeditational hypotheses adapted for structural equation modelling(Holmbeck, 1997; Shrout & Bolger, 2002) was employed to test formediation in the indirect effects. Accordingly, bootstrap-generatedbias-corrected confidence intervals were used to estimate thestandardized path coefficients representing the direct effects, withand without the inclusion of the mediating variables. Without themediating variables the direct path coefficients between maturityand physical self-worth (b ¼ �.30 (BBC 90% CI ¼ �.38, �.21),

Maturity Status

Sport Competence

SMC=.01

Physical ConditionSMC=.08

Body Attractiveness

SMC=.15

Strength SMC=.02

Physical Self-WorthSMC=.79

Physical Activity

SMC=.20

.45 (.07)

-.09 (.07)

-.29 (.07)

-.38 (.06)

.15 (.06)

.14 (.24)

.11 (.10)

.24 (.19)

.57 (.08)

Note. All solid line parameters are significant (p<.05). Dashed line parameters are non-significant. Standardized Beta coefficients are presented by each parameter with standard errors in parentheses. Indirect effects were observed between maturity status and physical self-worth (β = -.28 (BBC 90% CI = -.37, -.18), P<.01), and maturity status and physical activity (β = -.13 (BBC 90% CI = -.18, -.07), P<.01). Factors indicators are not included in the model for the purpose of making the presentation less complex. Correlations between the errors associated with the four sub-dimensions of the physical self were all significant and were as follows; sport competence & physical condition r=.91; sport competence & body attractiveness r=.61; sport competence & strength r=.65; physical condition & body attractiveness r=.57; physical condition & strength r=.49; body attractiveness & strength r=.32.

Fig. 1. Mediated effects model describing relations among biological maturity status, physical self-concept, and physical activity in adolescent British females.

Table 3Summary of regression analysis for variables predicting physical activity in ado-lescent British females.

Variable Step 1b(SE)a

Step 2b(SE)a

Step 3b(SE)a

Age �.15(.04)* �.07(.03) �.06(.03)

L. Jackson et al. / Psychology of Sport and Exercise 14 (2013) 447e454 451

p < .001), and maturity and PA (b ¼ �.23 (BBC 90% CI ¼ �.34, �.11),p < .01) were negative and statistically significant. However, whenthe mediating variables were included in the model the path co-efficients representing the direct effects between maturity andphysical self-worth (b ¼ �.06 (BBC 90% CI ¼ �.23, .08), p ¼ .37),maturity and PA (b ¼ �.11 (BBC 90% CI ¼ �.23, .01), p ¼ .14) wereattenuated and non-significant, yet not zero. Further, the inclusionof the direct effects between maturation and physical self-worth,Dc2(df ¼ 1) ¼ 1.2, p > .05, and maturation and PA,Dc2(df ¼ 1) ¼ 2.4, p > .05, did not result in statistically significantimprovements in model fit. As the coefficients associated with thedirect effects were not reduced to zero when the mediating vari-ables were included, the results suggest that the indirect effects ofmaturation on physical self-worth and PA were partially, yet notfully, mediated by the intervening variables (Preacher & Hayes,2004).

Inspection of the modification indices and standardized re-siduals associated with the hypothesized models did not indicateany improvements to the model that were theoretically or sub-stantively justified. Though the deletion of non-significant pathcoefficients would have resulted in superior model fit, we chose notto do this in order to better compare the hypothesized model rel-ative to previous versions of the model.

Maturity status �.10(.03) �.09(.03)Parental support for PA .51(.01)*** .51(.01)***Maturity status �

parental support.04(.01)

F 5.83* 35.23*** 26.54***R2 .02 .31 .31

*p < .05; **p < .01; ***p < .001.a Standardized Beta-coefficients and standard errors from the various steps of the

regression model.

Test of moderated effects

Results of the hierarchical regression analyses showing themaineffects of age, maturation, and parental support and the interactiveeffect of maturity and parental support on PA are presented inTable 3. Among participants, the final regression model explained

31% of the variance in PA. As predicted, parental support for PAserved as a positive and significant predictor of PA. Age, maturitystatus, and the interaction between maturity status and parentalsupport did not act as significant predictors of PA variance.

Discussion

The results of this study provided partial support for the Bio-cultural Model of Maturity-associated Variance in Adolescent PA(Cumming, Sherar, et al., 2011). Whereas physical self-conceptpartially mediated an inverse and indirect relation betweenmaturity status and PA, parental support for PA did not moderaterelations between maturation and PA. The hypothesized modeldescribing relations betweenmaturity status, physical self-concept,and PA, demonstrated adequate fit relative to the data. Further, the

L. Jackson et al. / Psychology of Sport and Exercise 14 (2013) 447e454452

magnitude and direction of the model’s path coefficients wereconsistent with those observed in previous versions of the model(Cumming, Standage, et al., 2011; Hunter Smart et al., 2012).Maturity status, though unrelated to perceptions of sport com-petence, inversely predicted perceptions of body attractiveness,physical condition, and positively predicted perceptions ofstrength. As with previous studies maturity status was most closelyassociated with perceptions of body attractiveness (Cumming,Standage, et al., 2011; Hunter Smart et al., 2012) and relations be-tween maturity status and the various dimensions of physical self-concept were small to moderate in size. Body attractiveness servedas the only significant predictor of physical self-worth. As withprevious models, the association between these constructs wasstrong, suggesting that perceptions of attractiveness play animportant role in how adolescent girls define their physical self-worth (Cumming, Standage, et al., 2011; Hunter Smart et al.,2012). In turn, physical self-worth positively predicted involve-ment in PA. As with previous models (Cumming, Standage, et al.,2011; Hunter Smart et al., 2012), the association between theseconstruct was moderate-to-large in magnitude, suggesting thatphysical self-worth contributes to the involvement of adolescentfemales in PA. Also consistent with previous models (Cumming,Standage, et al., 2011; Hunter Smart et al., 2012), inverse indirectrelations were observed between maturity and physical self-worth,and maturity and PA. The indirect associations between theseconstructs were, however, small to moderate in size.

The current study identifies a potential mechanism (i.e., physicalself-concept) through which maturity associated reductions in thePA of adolescent females may be explained and potentially coun-tered. Earlymaturing girls’ lower perceptions of physical self-worthand body attractiveness may reflect greater pubertal gains in bodysize and absolute and relative fat mass. These changes are lesscompatible with successful engagement in most forms of PA andless consistent with western ideals pertaining to attractiveness.Negative perceptions of the physical self and a lack of subjectiveand cognitive knowledge with regards to the female body and theprocesses of growth and maturation have been documented assources of anxiety in adolescent girls (Martin, 1996).

Those interested in promoting positive body image and healthbehaviours in adolescent females should design interventions toeducate students and, potentially, teachers and parents on theprocesses of growth and maturation; encouraging girls to viewpuberty as a normal and attractive consequence of becoming anadult and promoting acceptance of diversity in appearance. Suchinterventions have the potential to help girls adapt more positivelyto puberty, encouraging more positive engagement in health pro-moting behaviours. Interventions of this manner would requireeducation and the personal reorganization and reinterpretation ofbody image and related self-concepts. They might also extend fromexisting initiatives such as Cash’s body-imageCognitive BehaviouralTherapy (CBT) program (Cash, 1997). Although body-image CBTprograms are generally effective in collegiate and adult samples, fewstudies have examined the efficacy of such interventions in ado-lescents. Indeed, most interventions seeking to promote positivephysical self-concept in adolescents have limited their strategies tothe promotion of exercise (Fox, 2000). Though exercise can enhancephysical self-concept, the effectiveness of such interventions wouldappear somewhat limited in that they do not challenge the beliefsand perceptions that underpin physical self-concept.

The results of the regression model revealed a statistically sig-nificantmain effect of parental support on PA, but nomain effect formaturity status or an interaction effect for maturity and parentalsupport. This observation suggests that parental support for PAserves as a positive predictor of adolescent involvement in PA,irrespective of maturity status. That is, parental support for PA does

not appear to be any more or less important for girls of varyingmaturity status.

There are a number of possible reasons why parental support forPA did not moderate relations between maturation and PA. First,the influence of parental support varies with age, with markedreductions occurring during early and middle adolescence.Whereas some theorists view these reductions as a function ofincreased peer influence, others suggest that peer and parentallevels of influence operate independently (Newman & Newman,2008). Reductions in parental influence may occur quite earlywithin the context of PA, with adolescents becoming more auton-omous, more capable of directing and organizing their own play,and less dependent upon adults for logistical or emotional support.Consistent with this assumption, a one-way ANOVA by age groupdemonstrated that levels of perceived parental of support generallydeclined with age, F(3,240) ¼ 3.83, p < .05 (with the exception ofchildren aged 14 years see Table 1).

Another explanation as to why parental support did not mod-erate relations between maturation and PA is that parental supportand acceptance regarding the changes associated with growth andmaturity may be more important than it is in relation PA. Howparents interpret and react to pubertal changes in size, morphol-ogy, appearance, and function, and the extent to which parentsencourage their children to view and accept puberty as positive andnormal event (and not a barrier to PA) may play a more importantrole in terms of mitigating any negative effects of advancedmaturation. Summers-Efler (2004) noted that early maturing girlswho remained active through puberty reported high levels ofsupport and acceptance from parents, preventing the internal-ization or development of defensive strategies. For example, oneinterviewee commented that her father’s support and under-standing in relation to puberty served as a motivating factor for herto remain actively engaged in sports. “My dad was the soccer coach.He also tried to make me feel very comfortable. It let me know that heunderstood” (Summers-Efler, 2004, p. 39).

Although the results of this study did not support the concept ofmoderated effects, researchers should not be discouraged from thisline of investigation. As noted, support and acceptance pertaining tothe processes of growth andmaturationmay have greater relevanceto the relation between maturation and PA than support for PA, it-self. Peers, coaches, and PE teachers may also play more importantroles in moderating relations between these constructs. To betteridentify and understand sociocultural factors that contribute tomaturity associated variance in adolescent PA, researchers shouldemploy longitudinal andmixedmethods designs (i.e., combined useof quantitative and qualitative methods of analysis). This wouldenable researchers to identify different “trajectory groups”, for thepurpose of understanding how the groups differ as a function ofdevelopmental changes in relevant biological, individual, social, andenvironmental factors (Holmbeck, 2002, p. 413). Such approach(Brannen, 2005) could explain individual differences in the adjust-ment of early maturing females (i.e., why some remain active andothers do not). This knowledge could also inform subsequent in-terventions, helping practitioners more accurately identify those inneed of support and more effectively tailor PA interventions.

Limitations associated with the current study should be recog-nized. First, the results are limited to adolescent girls at a same sexschool in England. Relations among the variables of interest mayvary relative to age, ethnicity, social circumstance, and culture.Body attractiveness, for example, is considered a more salientsource of self-worth in working class girls (Martin, 1996), and Af-rican American girls tend to be more satisfied with their bodiesthan white girls (Parker et al., 1995). Second, inferences regardingcause and effect relations cannot be made due to the study’s cross-sectional design. To obtain more precise information about how

L. Jackson et al. / Psychology of Sport and Exercise 14 (2013) 447e454 453

physical self-perceptions might mediate relations between matu-ration and PA, longitudinal research is required, ideally beginningin late childhood and tracking changes through to late adolescence.Third, the method for estimating maturity status was derived fromdata collected in the United States. British and American girls aged9e15 years present very similar mean heights and weights (Cole,1994) (http://www.cdc.gov/growthcharts), however, furtherresearch is required to validate the formulae in British samples.Fourth, self-report measures of PA rely on the participant’s memoryto assess activity levels and can potentially be influenced by socialdesirability. Despite these limitations, the PAQ-A is considereda valid, low cost, and easy to use measures of activity that is wellsuited for use within youth (Kowalski et al., 2004). Finally, theparental support questionnaire did not differentiate between sup-port provided by fathers and mothers. It is possible that the impactof support on PAmay vary relative to sex of the parent or the degreeto which each parent is involved in their life.

In summary, this study observed that physical self-conceptpartially mediated an inverse relation between maturity statusand PA in adolescent females. Contrary to expectations, parentalsupport for PA did not mitigate the negative effects of withadvanced maturation of PA.

References

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting in-teractions. Newbury Park, CA: Sage Publications.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction insocial psychological research: conceptual, strategic and statistical consider-ations. Journal of Personality and Social Psychology, 51, 1173e1182. http://dx.doi.org/10.1037//0022-3514.51.6.1173.

Bayer, L. M., & Bayley, N. (1959). Growth diagnosis: Selected methods for interpretingand predicting development from one year to maturity. Chicago: University ofChicago Press.

Bradley, R. H., McRitchie, S., Houts, R. M., Nader, P., & O’Brien, M. (2011). Parentingand the decline of physical activity from age 9 to 15. International Journal ofBehavioral Nutrition and Physical Activity, 8(33). http://dx.doi.org/10.1186/1479-5868-8-33.

Brannen, J. (2005).Mixingmethods: theentryofqualitativeandquantitativeapproachesinto the research process. The International Journal of Social Research Methodology,Special Issue, 8(3), 173e185. http://dx.doi.org/10.1080/13645570500154642.

Byrne, B. (2001). Structural equation modeling with AMOS: Basic concepts, applica-tions, and programming. Mahwah, NJ: LEA.

Cash, T. F. (1997). The body-image workbook: An eight step program for learning to likeyour looks. Oakland, CA: New Harbinger.

Caspi, A., &Moffitt, T. E. (1991). Individual differences are accentuated during periodsof social change: the sample case of girls at puberty. Journal of Personality andSocial Psychology, 61, 157e168. http://dx.doi.org/10.1037/0022-3514.61.1.157.

Child Growth Foundation. (2005). Girls growth chart birth e 18 years. London: ChildGrowth Foundation.

Cole, T. J. (1994). Growth charts for both cross-sectional and longitudinal data.Statistics in Medicine, 13(23e24), 2477e2492. http://dx.doi.org/10.1002/sim.4780132311.

Cumming, S. P., Battista, R. A., Standage, M., Ewing, M. E., & Malina, R. M. (2006).Estimated maturity status and perceptions of adult autonomy support in youthsoccer players. Journal of Sports Sciences, 24(10), 1039e1046. http://dx.doi.org/10.1080/02640410500386142.

Cumming, S. P., Sherar, L. B., Pindus, D. M., Coelho e Silva, M. J., Malina, R. M., &Jardine, P. R. (2011). A biocultural model of maturity associated variance inadolescent physical activity. International Review of Sport and Exercise Psychol-ogy, 5, 23e43. http://dx.doi.org/10.1080/1750984X.2011.630481.

Cumming, S. P., Standage, M., Loney, T., Gammon, C., Neville, H., Sherar, L. B., et al.(2011). The mediation effect of physical self-concept on relations betweenbiological maturity status and physical activity in adolescent females. Journal ofAdolescence, 34, 465e473. http://dx.doi.org/10.1016/j.adolescence.2010.06.006.

Davison, K. K., Werder, J. L., Trost, S. G., Baker, B. L., & Birch, L. L. (2007). Why are earlymaturing girls less active? Links between pubertal development, psychologicalwell-being, and physical activity among girls at ages 11 and 13. Social Science &Medicine, 64, 2391e2404. http://dx.doi.org/10.1016/j.socscimed.2007.02.033.

Dollman, J., Norton, K., & Norton, L. (2005). Evidence for secular trends in children’sphysical activity behaviour. British Journal of Sports Medicine, 39(12), 892e897.http://dx.doi.org/10.1136/bjsm.2004.016675.

Eisenmann, J. C., & Wickel, E. E. (2009). Biology of physical activity in children:revisited. Pediatric Exercise Science, 21(3), 257e272.

Eklund, R. C., Whitehead, J. R., & Welk, G. J. (1997). Validity of the children andyouth physical self-perception profile: a confirmatory factor analysis. ResearchQuarterly for Exercise and Sport, 68(3), 249e256.

Epstein, L., Valoski, A. M., Kalarchian, M. A., & McCurley, J. (1995). Do children loseand maintain weight easier than adults? A comparison of child and parentweight changes from six months to ten years. Obesity Research, 3, 411e417.

Fox, K. R. (2000). Self-esteem, self-perceptions and exercise. International Journal ofSport Psychology, 31, 228e240.

Harter, S. (1981). A model of intrinsic mastery motivation in children: Individual dif-ferences and developmental change, Vol. 14. New Jersey: Hillsdale.

Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity inthe study of mediators and moderators: examples from the child-clinical andpediatric psychology literatures. Journal of Consulting and Clinical Psychology,65(4), 599e610.

Holmbeck, G. N. (2002). A developmental perspective on adolescent health andillness: an introduction to the special issues. Journal of Pediatric Psychology,27(5), 409e416. http://dx.doi.org/10.1093/jpepsy/27.5.409.

Hu, L., & Bentler, P. (1999). Cut off criteria for fit indexes in covariance structureanalysis: conventional criteria versus new alternatives. Structural EquationModeling, 6, 1e55. http://dx.doi.org/10.1080/10705519909540118.

Hunter Smart, J. E., Cumming, S. P., Sherar, L. B., Standage, M., Neville, H., &Malina, R. M. (2012). Maturity associated variance in physical activity andhealth-related quality of life in adolescent females. A mediated effects model.Journal of Physical Activity & Health, 9, 86e95.

van Jaarsveld, C. H. M., Fidler, J. A., Simon, A. E., & Wardle, J. (2007). Persistentimpact of pubertal timing on trends in smoking, food choice, activity, and stressin adolescence. Psychosomatic Medicine, 69(8), 798e806. http://dx.doi.org/10.1097/PSY.0b013e3181576106.

Khamis, H. J., & Roche, A. F. (1994). Predicting adult height without using skeletalage: the Khamis-Roche method. Pediatrics, 94, 504e507, (Pediatrics, 595, 457,1995 for the corrected version of the tables).

Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.).New York: Guilford.

Kowalski, K. C., Crocker, P. R. E., & Donen, R. M. (2004). The physical activity ques-tionnaire for older children (PAQ-C) and adolescents (PAQ-A) manual. Saskatoon,CA: University of Saskatchewan.

Kowalski, K. C., Crocker, P. R. E., & Kowalski, N. P. (1997). Convergent validity of thephysicalactivityquestionnaire foradolescents.PediatricExerciseScience, 9, 342e352.

Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or notto parcel: exploring the question, weighing the merits. Structural EquationModeling, 9(2), 151e173. http://dx.doi.org/10.1207/S15328007SEM0902_1.

MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidence limits for theindirect effect: distribution of the product and resampling methods. Multi-variate Behavioral Research, 39(1), 99e128. http://dx.doi.org/10.1207/s15327906mbr3901_4.

Malina, R. M. (1995). Anthropometry. In P. J. Maud, & C. Foster (Eds.), Physiologicalassessment of human fitness (1st ed.). (pp. 205e219) Champaign, IL: HumanKinetics.

Malina, R. M. (2008). Biocultural factors in developing physical activity levels. InA. L. Smith, & S. J. H. Biddle (Eds.), Youth physical activity and inactivity (pp. 141e166). Champaign, IL: Human Kinetics.

Malina, R. M., Bouchard, C., & Bar-Or, O. (2004). Growth maturation and physicalactivity. Champaign, IL: Human Kinetics.

Malina, R. M., Dompier, T. P., Powell, J. W., Barron, M. J., & Moore, M. T. (2007).Validation of a noninvasive maturity estimate relative to skeletal age in youthfootball players. Clinical Journal of Sport Medicine, 17(5), 362e368. http://dx.doi.org/10.1097/JSM.0b013e31815400f4.

Malina, R. M., Morano, P. J., Barron, M., Miller, S. J., & Cumming, S. P. (2005). Growthstatus and estimated growth rate of youth football players: a community-basedstudy. Clinical Journal of Sport Medicine, 15(3), 125e132. http://dx.doi.org/10.1097/01.jsm.0000164287.42066.63.

Martin, K. A. (1996). Puberty, sexuality, and the self: Girls and boys at adolescence.New York: Routledge.

Mrug, S., Elliott, M., Gilliland, M. J., Grunbaum, J. A., Tortolero, S. R., Cuccaro, P., et al.(2008). Positive parenting and early puberty in girls e protective effects againstaggressive behavior. Archives of Pediatrics & Adolescent Medicine, 162(8), 781e786. http://dx.doi.org/10.1001/archpedi.162.8.781.

Newman, B. M., & Newman, P. R. (2008). Development through life: A psychosocialapproach (10th ed.). Belmont, CA: Wadsworth Cengage Learning.

Niven, A. G., Fawkner, S. G., Knowles, A., & Stephenson, C. (2007). Maturationaldifferences in physical self-perceptions and the relationship with physical ac-tivity in early adolescent girls. Pediatric Exercise Science, 19, 472e480.

Parker, S., Nichter, M., Nichter, M., Vuckovic, N., Sims, C., & Ritenbaugh, C. (1995).Body-image and weight concerns among African-American and white adoles-cent females e differences that make a difference. Human Organization, 54(2),103e114.

Petersen, A. C., & Taylor, B. (1980). The biological approach to adolescence: bio-logical change and psychological adaptation. In J. Adelson (Ed.), Handbook ofadolescent psychology (pp. 117e155). New York: Wiley-Interscience.

Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirecteffects in simple mediation models. Behavior Research Methods Instruments &Computers, 36(4), 717e731.

Riddoch, C., Mattocks, C., Deere, K., Saunders, J., Kirkby, J., Tilling, K., et al. (2007).Objective measurement of levels and patterns of physical activity. Archives ofDisease in Childhood, 92, 963e969. http://dx.doi.org/10.1136/adc.2006.112136.

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation ofintrinsic motivation, social development, and well-being. American Psychologist,55(1), 68e78.

L. Jackson et al. / Psychology of Sport and Exercise 14 (2013) 447e454454

Sallis, J. F. (2000). Age-related decline in physical activity: a synthesis of human andanimal studies. Medicine and Science in Sports and Exercise, 32(9), 1598e1600.

Sherar, L. B., Cumming, S. P., Eisenmann, J. C., Baxter-Jones, A. D. G., & Malina, R. M.(2010). Adolescent biological maturity and physical activity: biology meetsbehaviour. Pediatric Exercise Science, 22(3), 332e349.

Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimentalstudies: new procedures and recommendations. Psychological Methods, 7, 422e445. http://dx.doi.org/10.1037//1082-989X.7.4.422.

Suhr, D. (2006, 11/07/2012). The basics of structural equation modeling. Retrievedfrom. http://www.lexjansen.com/wuss/2006/tutorials/TUT-Suhr.pdf.

Summers-Efler, E. (2004). Little girls in women’s bodies: social interaction and thestrategizing of early breast development. Sex Roles, 51, 29e44. http://dx.doi.org/10.1023/B: SERS.0000032307.16204.ec.

Sweet, S. L., Dompier, T. P., Stoneberg, K. N., & Ragan, B. G. (2002). Self-reportedparent stature is acceptable in estimates of maturity status in youth soccerplayers. Journal of Athletic Training, 4(2), S-129.

Trost, S. G., Sallis, J. F., Pate, R. R., Freedson, P. S., Taylor, W. C., & Dowda, M. (2003).Evaluating a model of parental influence on youth physical activity. AmericanJournal of Preventive Medicine, 25(4), 277e282. http://dx.doi.org/10.1016/S0749-3797(03)00217-4.

Whitehead, J. R. (1995). A study of children’s physical self-perceptions using anadapted physical self-perception profile questionnaire. Pediatric Exercise Sci-ence, 7, 132e150.

Wickel, E. E., & Eisenmann, J. C. (2007). Maturity-related differences in physicalactivity among 13e14-year old adolescents. Pediatric Exercise Science, 19(4),384e392.