Research ArticleAdolescent Sports Behavior and Social Networks The Role ofSocial Efficacy and Self-Presentation in Sports Behavior

Lei Lei12 Huifang Zhang 2 and Xin Wang3

1Sports Department Northwest AampF University Yangling 712100 China2School of Economics and Finance Xirsquoan Jiaotong University Xirsquoan 710064 China3Development Planning Office Beijing Institute of Computer Technology and Application Beijing 100854 China

Correspondence should be addressed to Huifang Zhang zhanghuifangxjtueducn

Received 18 June 2020 Revised 16 July 2020 Accepted 21 July 2020 Published 29 August 2020

Guest Editor Liang Wang

Copyright copy 2020 Lei Lei et al +is is an open access article distributed under the Creative Commons Attribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Social networks are a complex system that members communicate create new connections or destroy existing connections andfurther deliver major impacts on each memberrsquos life Given the spread of the Internet and increased academic pressure sedentaryand prescreen behaviors are very common among adolescents meanwhile sports behaviors are gradually decreasing +issituation has had an adverse effect on health+is paper used a questionnaire survey to investigate the influence of social networkson adolescent sports behavior including the intermediary role of social efficacy and moderating effect of self-presentation +equestionnaire survey was conducted on 568 students from 6 high schools in Shaanxi Henan and Shandong Provinces After thisfactor analysis and weighted least squares method were used for the empirical test Based on theoretical and empirical analysis thispaper found the following (1) Social networks of adolescents have obvious positive predictive effects on their sports behavior Asingle online social network and an offline social network instrumental network emotional network and mixed network haveobvious positive predictive effects on adolescent sports behaviors However under the influence of multiple types of socialnetworks an offline social network has a negative predictive effect while a mixed network has effects that are not as obvious (2)Social efficacy plays an intermediary role in the relationship between social network and adolescent sports behavior (3) +emoderating effect of self-presentation is not significant

1 Introduction

Today nearly four-fifths of the worldrsquos adolescents lackenough physical activity and the situation for Chinese ad-olescents is even worse Sedentary behavior and prescreenbehavior are very common in adolescentsrsquo study and daily lifeactivities In 2019 the overweight and obesity rates of Chineseyoung students were 285 and 117 respectively +emyopia or short sight rate for primary school students middleschool students and college students is also increasingcurrently ranging from 457 to 877 Adolescents are thefuture of the world and they are a group that needs specialattention Active participation in physical activities is con-ducive to the development of both the physical and mentalhealth of adolescents and it is thus an urgent task to promotegreater adolescent physical activity levels

In view of the basic trend of social development which ischanging from relatively closed small group to increasinglyopen networked community life human society has reachedan unprecedented level in terms of network structurenetwork connection density and network member com-munication Social networking not only is a static rela-tionship structure but also offers a variety of dynamics forthe now microprocess of cognition and interpersonal in-teraction To pay close attention to the complex unique andcontinuous interaction between social network and actors itis necessary to discuss the sociability of actors and theirembedded network relationships Although the influenceand intervention of social networking on health risk be-haviors such as smoking alcohol abuse and sexual behaviorof adolescents have been extensively studied the changesand ongoing development of adolescentsrsquo physical activity

HindawiComplexityVolume 2020 Article ID 4938161 10 pageshttpsdoiorg10115520204938161

behavior a complex social problem need to be furtherstudied at the level of social networking for the most relevantfactors theories and mechanisms

Compared to early childhood and adulthood adoles-cents are more influenced by their peers during their processof socialization [1 2] Studies have found that social net-working can not only promote individualsrsquo participation incollective actions and ldquocatalyzerdquo large-scale collective actionsbut also establish cross-organizational connections throughits networks [3] Making friendship with others can increaseadolescentsrsquo motivation for sports For example distancecycling in front of friends or in a group will be greater inlength than when they are cycling alone indicating thatonersquos peers do enhance the enthusiasm for sports activities[4] Indeed if adolescents have more friends and a largerproportion of same-sex friends they will participate in morephysical activities after school +e intensity of physicalactivity during recess also positively correlates with thenumber of friends who participate in sports +is situationalso varies according to gender for example boys are moreeasily affected by their network of friends than girls are [5]In addition researchers believe that adolescentsrsquo physicalactivity behavior is influenced by their siblings and friendsFriends have a greater influence on adolescent participationin organized physical activities while siblings mainly affectinformal and spontaneous physical activity participationBesides the supports of teachers also influence adolescentsfor teachers and adolescents spending most time together instudying Sometimes the teacher-students relationship isimportant in adolescentsrsquo behavior because of regulationsand commands At the same time parental support is crucialbecause it can reduce the internal perception barriers ofadolescents and affect their daily participation in physicalactivities If parents were and are also active participants inphysical activity their adolescent children will be more likelyto show active physical activity behavior as well [6]

Although the social network environment should beconsidered for any intervention of adolescentsrsquo physicalactivities it is worth noting that the influence of friends andpartners on adolescentsrsquo physical activities is not regulatedby their knowledge of those physical activities but instead bythe interdependent relationship between the network se-lected by the peers and their physical activity [7] +ereforein any intervention of physical activity behaviors of ado-lescents onersquos choice of friends may be critical for pro-moting and maintaining health and positive behaviors Inother words social networkingmay have a positive influenceon adolescent sports behavior

Given the popularization of computers network plat-forms have gradually become the main channel for ado-lescents to use to reach out and make friends Adolescentsnot only form offline social network but also form onlinesocial networks Adolescents who are interested in physicalactivities online prefer to be closer to relevant topics and aremore likely to provide information about their physicalactivities [8] +at is to say given a digital background notonly may the offline social network actively promote ado-lescent sports behavior but it also may promote individualsports behavior as well

In reality the theoretical model of ldquotwo paths and threelayersrdquo of individual behavior suggests that stable personalitytraits and external situations affect individual behaviors viainternal cognitive processes Moreover social cognitivetheory points out that individual behavior subject cognitionand environment interact dynamically while individualcognition is the core of such interaction [9] Both theoriesdemonstrate that internal cognition plays an importantintermediary role between external situations and personalbehavior Bandura pointed out that a sense of efficacy is animportant cognitive factor that not only affects an indi-vidualrsquos choice of behavior but also determines the effortlevel and the ability to overcome obstacles [10]

Since self-efficacy is domain-specific Fan and Mak putforward social efficacy [11] to show that there is an obviouspositive correlation between social relations and social ef-ficacy [12] Individuals with low social efficacy tend to in-terpret uncertain social situations as dangerous and havenegative reactions including excessive self-concern anxietyand operational behavior all obstacles used to avoid andwithdraw from real social situations [13] and further reduceindividual behaviors To sum up then in a real-life com-munication situation social networks social efficacy andindividual behavior are closely related Indeed social efficacymay play an intermediary role in the social network role ininfluencing adolescent sports behavior

+e popularity of computers and the diversification ofnetwork platforms now enable adolescents to present them-selves at any time and share and learn on network platformssuch as Facebook or WeChat With the rise of sports apps andquantitative equipment many adolescents now use sportssoftware to exercise and show and share themselves on socialplatforms For example the development of Keep and otherfitness apps allows adolescents to not only learn the existingsports videos but also share their sports videos and sportsrecords on the same platform and promote their continuoussports activity On the one hand this development may raiseadolescentsrsquo attention toward their own health and strengthentheir sports behavior on the other hand it can also make otheradolescents undertake sports through the communication andsharing of them on social networks +us the adolescentsrsquo self-presentation on network platforms is likely to strengthen theinfluence of social networking on adolescent sports behavior+ese above thoughts mean that diverse networks may pro-mote adolescentsrsquo self-presentation and that self-presentationmay encourage adolescents to do sportsmore frequently whichshows that self-presentation indeed has a moderating effect

+rough undertaking the above analysis this paperexamines the influence of social networking on adolescentsports behavior the mediating effect of adolescentsrsquo socialefficacy and the moderating effect of self-presentationBased on social cognition theory this paper proposes amodel that includes a mediating effect and a moderatingeffect and puts forward three hypotheses H1 social net-works positively promote adolescent sports behavior H2social efficacy plays a mediating role in the relationshipbetween social network and adolescent sports behavior andH3 self-presentation plays a moderating role for socialnetworking and adolescent sports behavior

2 Complexity

At the same time to deeply analyze the influence ofdifferent social networks on adolescents sports this studydivides social networks into online and offline social plat-forms and emotional instrumental and mixed social net-works Of these teachers and counselors belong to theinstrumental network family members and lovers belong tothe emotional network and roommates and classmatesbelong to a mixed network [11] +rough an analysis ofsocial networks social efficacy and self-presentation thispaper offers both a theoretical basis and practical guidancefor promoting the establishment of social networks andfacilitating adolescent sports behavior

2 Materials and Methodology

21 Subjects +e subjects were 630 adolescents from threejunior middle schools and three senior high schools inShaanxi Henan and Shandong Provinces A total of 630questionnaires were distributed and of these 568 validquestionnaires were collected providing an effective re-covery rate of 901 +ere were 296 boys (5211) and 272girls (4789) 376 senior students and 192 junior students+e average age of the subjects was 1452 years (SD 168with the age range being 12ndash18)

22 Measurements An adolescent sports behavior scale isbased on the questionnaire by Mao et al [14] which in-cluded 6 items (eg I usually take exercise with friends Ihave the habit of exercising) and the questionnaire wasscored using 5 answers (1 ldquototally inconsistentrdquo to 5ldquocompletely consistentrdquo) In this study the confirmatoryfactor analysis of the questionnaire fits well χ2df 284RMSEA 005 NFI 0995 GFI 0998 and CFI 0997+e internal concordance coefficient α was 0687

+e social network scale uses a questionnaire designedby Park et al [15] which contains 14 items and is dividedinto either two dimensions as (1) an online social network (6items eg Many of my friends in real life love sports) and (2)an offline social network (4 items eg +e number ofcommunities Irsquove join online) or three dimensions as (1) aninstrumental social network (4 items eg I have a closerelationship with my class teacher) (2) an emotional net-work (4 items eg I have a close relationship with myrelatives of the same age) and (3) a mixed network (4 itemseg I have a close relationship with my friends in real life)

Some of the items in the questionnaire on different di-mensions were the same For example ldquoI have a close rela-tionship with my friends in real liferdquo is a measurement item ofboth an offline social network and a mixed network +equestionnaire is scored using 5 points (1 ldquototally inconsistentrdquoto 5 ldquocompletely consistentrdquo) In the current study theconfirmatory factor analysis of the questionnaire fits wellχ2df 284 RMSEA 006 NFI 0998 GFI 0999 andCFI 0999+e internal concordance coefficient αwas 0654+e coefficients α of each dimension were 0940 (online socialnetwork) 0936 (offline social network) 0898 (instrumentalsocial network) 0766 (emotional social network) and 0847(mixed social network)

+e social efficacy scale used a questionnaire compiledby Jeong and Kim [16] +ere were 6 items on this ques-tionnaire in the study (eg online communication I caneasily become friends with other people I can easily talk tounfamiliar people) +e questionnaire was scored using 5points (1 ldquototally inconsistent ldquoto 5rdquo completely consistentrdquo)In this study the confirmatory factor analysis of the ques-tionnaire fits well χ2df 124 RMSEA 002 NFI 0998GFI 0999 and CFI 0999 +e internal concordancecoefficient α was 0634

+e self-presentation scale used a questionnaire com-piled by Kim and Lee [17] It included 4 items (eg I will postphotos that show the real me I donrsquot mind sharing some badthings that happened to me on the Internet) +e ques-tionnaire was scored based on 5 points (1 ldquototally incon-sistentrdquo to 5 ldquocompletely consistentrdquo) In this study theconfirmatory factor analysis of the questionnaire fits wellχ2df 224 RMSEA 002 NFI 0998 GFI 0999 andCFI 0999 +e internal concordance coefficient α was0820

23 Programs and Data Processing In this current studyafter obtaining the consent of the school leaders teachersand students the questionnaire practice was explained byhighly trained surveyors in accordance with standardizedinstruction and all questionnaires were collected immedi-ately SPSS 220 software was used to analyze the data andPROCESS V30 was used to test the mediating effect

First we used the factor analysis method to reduce thedimensionality of variables +e mathematical model was

x1 a11f1 + a12f2 + a13f3 + middot middot middot + a1kfk + ε1

x2 a21f1 + a22f2 + a23f3 + middot middot middot + a2kfk + ε2

x3 a31f1 + a32f2 + a33f3 + middot middot middot + a3kfk + ε3

xp ap1f1 + ap2f2 + ap3f3 + middot middot middot + apkfk + εp

⎧⎪⎪⎪⎪⎪⎪⎪⎪⎨

⎪⎪⎪⎪⎪⎪⎪⎪⎩

(1)

For equation (1) xi is the standardized variable fi is thefactor variable ε is the special factor aij is the factor loadkltp A matrix was used to simplify equation (1) as follows

X AF + ε (2)

where F is the factor variable matrix and A is a factor loadmatrix and satisfies cov(F ε) 0 D(F) Im D(ε) δi

+e factor load aij represents the degree of correlationbetween Xi and Fi and the square sum of the elements in thei-th row of the factor load matrix where A is h2

i 1113936kj1 a2

ij which represents the explanatory power of all the factorvariables for the total variance of Xi and Sj 1113936

p

i1 a2ij is the

variance contribution of the variable FjAfter determining the factor variables the factor load

matrix needed to be estimated If we set the eigenvalues ofthe sample covariance matrix and the corresponding stan-dard orthogonalized eigenvectors as λ1 ge λ2 ge middot middot middot ge λp ge 0e1 e2 ep then the covariance matrix can be decomposedinto

Complexity 3

1113944 U

λ1 0

λ2

0 λp

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝

⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠Uprime 1113944

p

i1λieieiprime (3)

When the last several eigenvalues are small the co-variance matrix can be approximately decomposed into

1113944 asymp

λ11113969

e1

λm

1113969

em1113874 1113875

λ1

1113968e1prime

λm

1113968emprime

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦+

σ21σ22

σ2p

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝

⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

AAprime + 1113944ε

(4)

where A is the factor covariance matrix Since the factorloading matrix is not unique factor rotation was executed tomake the meaning of the common factor clearer Consid-ering the orthogonal rotation of two factors the orthogonalmatrix then became

Q cos ϕ minussinϕ

sinϕ cosϕ11138891113888 (5)

and set

B AQ bij11138731113872 i 1 2 p j 1 2 (6)

where B is the rotation factor load matrix At this time werequired the variance of the two columnsrsquo data in B to be aslarge as possible which also meant that the relative varianceVi should also be as large as possible

Vj 1p

1113944

p

i1

b2ij

h2i

11138891113888

2

minus 1113944

p

i1

b2ij

h2i

11138891113888

2

j 1 2 (7)

Making dVdϕ 0 then ϕ should satisfy the followingequation

tan 4ϕ D0 minus 2A0B0( 1113857p( 1113857

C0 minus A20 minus B2

0( 1113857p( 1113857 (8)

where

A0 1113936p

i1ui B0 1113936

p

i1vi

C0 1113936p

i1u2

i minus v2i( 1113857 D0 2 1113936p

i1uivi

ui a2

i1hi

1113874 11138752

minusa2

i2hi

1113874 11138752 vi

2ai1ai2

h2i

⎧⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎨

⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎩

(9)

Using SPSS for the factor analysis taking adolescentsports and social network as examples the scree plots wereobtained It can be seen from Figure 1 that the adolescent

sports scale can extract one factor while the social networkscale can extract two factors

3 Results and Discussion

31 Descriptive Analysis +e mean standard deviation andthe correlation coefficient of each variable are shown inTable 1 +ese results show that there are obvious correla-tions between adolescent sports behavior and different socialnetworks and also obvious correlations between social ef-ficacy self-presentation social networking and sports Inaddition social networks especially offline social networkshave significant age differences Different types of socialnetworks and self-presentations also have gender differ-ences In order to explore the independent effects of socialnetwork on adolescent sports age and gender were used hereas the control variables

Furthermore we use histogram to basically analyze thesample From Figure 2(a) it can be seen that the sample isbalanced and distributed between male and female indi-cating the sample can reflect reality well In different gendergroup (Figure 2(b)) the number of adolescents who are 13-14 years old is the most while the number of 12 years oldadolescents is the least

From Figure 2(c) during the group of 12 years oldadolescents the scores of five social networks are all highindicating that for 12-year-olds their social network typedensity and intensity are stronger For 13-14 years oldadolescents all types of social networks are not high es-pecially online social network Its average value is 12 whichmeans the online social network is not strong +is maybecause the academic pressure for adolescents in the stage ishigher than other groups For 15-16 years old adolescentsmixed social network and emotional social network arestronger than other social networks and for 17-18 years oldadolescents online social network mixed social networkand emotional social network are stronger than other twosocial networks From Figure 2(d) no matter how old ad-olescents are their sports activities are frequent For ex-ample for 13-14 years old adolescents most adolescents dosports more than once a week only 5 of adolescents dosports once a week or even less than once a week FromFigure 2 we can clearly know the distribution of sample thesocial networks of different group and the sports behavior ofdifferent groups it helps to lay the foundation for the fol-lowing empirical research

32 Results

321 Influence of Social Networking on Adolescent SportsBehavior Econometric models are constructed to test therelationships between different social networks and ado-lescent sports behavior

Sporti βXi + PQ + e1 (10)

Among them Sport is adolescent sports behavior X isan adolescentsrsquo social network that includes an onlineofflinesocial network and an instrumentalemotionalmixed social

4 Complexity

network Q is the control variable matrix and e1 is theregression residual +e empirical results are shown inTable 2

+e results in Table 2 indicate that adolescentsrsquo socialnetworks have a significant positive effect on their sportsbehavior (β 0860 plt 0001) +e wider the homoge-neity of these adolescentsrsquo social relationship is the morefrequent their sports behavior becomes Online socialnetworks and offline social networks have positive effectson adolescent sports behavior respectively Howeverwhen adolescents have both an offline social network andan online social network the offline social network has apositive effect on adolescent sports (β 1381 plt 0001)while the online social network has a negative effect onadolescent sports (β 0517 plt 0001) An instrumentalnetwork emotional network and mixed networkhave positive effects on adolescent sports respectivelywith coefficients of 0735 0964 0880 and are significantat a 1 level However when adolescents have the threekinds of above social relations at the same time theemotional network and the mixed network still havepositive effects on adolescent sports but the influence ofthe instrumental network on adolescent sports then be-comes insignificant

322 Mediating Effect of Social Efficacy According to Wenand Ye [18] it is necessary to test the parameters of threeregression equations to verify the mediating effect

Sporti cSNi + PQ + e2 (11)

SEi αSNi + PQ + e3 (12)

Sporti cprimeSPi + bSEi + PQ + e4 (13)

Among these SN is social network SE is social efficacyQ is the control variable matrix and e2 sim e4 are the re-gression residuals Further c is the total effect of the in-dependent variable (social network) on the dependentvariable (adolescent sports behavior) a is the effect of socialnetwork on the intervening variable (social efficacy) b is theeffect of social efficacy on adolescent sports behavior aftercontrolling for the influence of social network cprime is the directeffect of social network on adolescent sports behavior aftercontrolling for the influence of social efficacy +e mediatingeffect is tested in five steps first test the coefficient c ofequation (11) if c is significant the intermediary effect issignificant otherwise there is a masking effect But whetherit is significant or not follow-up tests are carried out

1 2 3 4Number

Scree plot

000

050

100

150

200

250

300

350

Eige

nval

ues

(a)

1 2 3 4 5 6 7 8 9 10 11Number

Scree plot

000100200300400500600700800

Eige

nval

ues

(b)

Figure 1 Scree plot of adolescent sports (a) and social network (b)

Table 1 Mean standard deviation and correlation coefficient of variables

M SD 1 2 3 4 5 6 7 8 9 10 111 Age 1453 169 12 Gender 053 050 minus0018 13 Sports 288 093 minus0042 minus0076 14 Social network 208 087 minus0004lowast minus0105lowast 0858lowast 15 Online social network 209 089 0015lowast minus0084 0744lowastlowast 0967lowastlowast 16 Offline social network 236 064 minus0015 minus0113lowast 0903lowastlowast 0988lowastlowast 0916lowastlowast 17 Mixes social network 230 071 minus0018 minus0126lowastlowast 0876lowastlowast 0945lowast 0863lowastlowast 0966lowastlowast 18 Instrumental socialnetwork 234 066 0000 minus0107lowast 0734lowastlowast 0944lowastlowast 0923lowastlowast 0924lowastlowast 0823lowastlowast 1

9 Emotional socialnetwork 1453 169 minus0033 minus0066 0965lowastlowast 0844lowastlowast 0724lowastlowast 0891lowastlowast 0850lowastlowast 0711lowastlowast 1

10 Social efficacy 241 071 minus0058 minus0078 0980lowastlowast 0838lowastlowast 0725lowastlowast 0884lowastlowast 0880lowastlowast 0712lowastlowast 0916lowastlowast 111 Self-presentation 243 096 0006 minus0116lowastlowast 0808lowastlowast 0810lowastlowast 0710lowastlowast 0848lowastlowast 0895lowastlowast 0662lowastlowast 0816lowastlowast 0780lowastlowast 1Note N 568 gender is a dummy variable female 0 and male 1 lowastplt 005 lowastlowastplt 001

Complexity 5

Second test the coefficient a of equation (12) and the co-efficient b of equation (13) in turn if both are significantthen the indirect effect is significant go to the fourth step ifat least one is not significant go to the third step +ird useBootstrap method to directly check H0 ab 0 If it issignificant then the indirect effect is significant go to thefourth step otherwise the indirect effect is not significantstop the analysis Forth examining the coefficient cprime ofequation (13) if it is not significant the direct effect is not

significant indicating that there is only an intermediaryeffect If it is significant that is the direct effect is significantgo to the fifth step Fifth compare the signs of ab and cprime Ifthey are the same it is partly intermediary effect If the sign isdifferent it is a masking effect If a b and c are all significantthe mediation effect is significant otherwise there are othereffects For example if c is not significant there is a maskingeffect if cprime is significant there also exist direct effects Besidesthe above five steps a nonparametric percentile Bootstrap

Below 12 13-14 15-16 17-18

Statistics of gender in different ages

MaleFemale

010203040506070

()

(a)

Male Female

Statistics of ages in different genders

Below 1213-14

15-1617-18

05

101520253035404550

()

(b)

Online social networkOffline social networkMixed social network

Instrumental social networkEmotional social network

Below 12 13-14 15-16 17-18

Statistics of social networks in different ages

0

1

2

3

4

5

6

(c)

Once a weekOne to three times a weekFour to five times a weekMore than five times a week

Below 12 13-14 15-16 17-18

Statistics of sport behaviors in different ages

05

10152025303540

()

(d)

Figure 2 Statistic of adolescent social networks and sports behavior

Table 2 Estimated results of the influence of social network on adolescent sports behavior

Variables Adolescent sports behaviorConstant 0316 (1568) 0471lowast (1794) 0210 (1240) 0069 (0465) 0357 (1336) 0100 (0966) 0179 (0945) 0102 (1071)Age minus0023lowast (minus1667) minus0031lowast (-1767) minus0016 (minus1427) minus0007 (minus0737) minus0025 (minus1366) minus0006 (minus0859) minus0015 (minus1157) minus0007 (minus1113)Gender 0028 (0599) minus0028 (minus0458) 0053 (1356) 0074lowastlowast (2154) 0004 (0068) minus0024 (minus1004) 0070 (1596) 0004 (0203)Socialnetwork 0860lowastlowastlowast (37052) mdash mdash mdash mdash mdash mdash mdash

Online socialnetwork mdash 0744lowastlowastlowast (24673) mdash minus0517lowastlowastlowast (minus12194) mdash mdash mdash mdash

Offline socialnetwork mdash mdash 0905lowastlowastlowast (45545) 1381lowastlowastlowast (32455) mdash mdash mdash mdash

Instrumentalnetwork mdash mdash mdash mdash 0735lowastlowastlowast (23892) mdash mdash 0016 (0818)

Emotionalnetwork mdash mdash mdash mdash mdash 0964lowastlowastlowast (81163) mdash 0793lowastlowastlowast (38244)

Mixednetwork mdash mdash mdash mdash mdash mdash 0880lowastlowastlowast (40217) 0189lowastlowastlowast (7352)

Adj-R2 0737 0554 0816 0858 0538 0931 0767 0942

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

6 Complexity

method was used to test the mediating effect If the 95confidence interval does not contain a zero then the me-diating effect becomes significant +e empirical regressionresults and significance test results for the mediating effectare shown in Table 3

+e results in Table 3 show that R2 in the three equationswas bigger than 07 and the goodness of fit for the model wasgood In equation (11) social network has a significantpositive effect on adolescent sports behaviorc 0860 plt 0001 In equation (12) social network has asignificant effect on social efficacy c 0839 plt 0001 Inequation (13) social network and social efficacy have asignificant effect on adolescent sports behaviorcprime 0122 b 0879 plt 0001 According to the mediatingeffect test procedure of Wen and Ye [18] the mediatingeffect of social efficacy is significant because a b c and cprime areall significant Further the Bootstrap method was used totest the mediating effect of social efficacy 5000 Bootstrapsamples were randomly selected from the original sample forindirect effect estimation Table 3 shows the estimated valueof the indirect effect of social efficacy to be 07374 and 95confidence intervals were [07007 07704] excluding zero+us the mediating effect of social efficacy becamesignificant

323 Moderating Effect of Self-Presentation As for themoderating effect of self-presentation the followingeconometric model was constructed

Sporti αSNi + βSPi + cSNi lowast SPi + PQ + e5 (14)

where SP is the self-presentation +e stepwise regressionmethod was used to estimate the moderating effect of self-presentation +e results are shown in Table 4

+e results of the study show that although both socialnetworking and self-presentation have significant positiveeffects on adolescent sports behavior the interaction termsof social network and self-presentation do not have anysignificant influence on adolescent sports behavior therebyindicating that the moderating effect of self-presentation isnot significant

In addition this study found that age has a negativepredictive effect on adolescent sports behavior that is theolder the adolescents are the less likely they are to exerciseGender had no significant effect on the prediction of ado-lescent sports behavior

33 Discussion

331 Influence of Social Networks on Adolescent SportsBehavior +e results of this study indicate that socialnetworks have significant influences on adolescent sportsbehavior demonstrating that adolescents with larger anddenser social networks have more frequent sports behaviorPeople are mutually linked to each other and so are theirsports behaviors +us the existence of social networksmeans that personsrsquo movements are interdependent [19]Sports behavior can be spread from person to person and

there is also a phenomenon called ldquopeer effectsrdquo [20] whichis just as the proverb says ldquoIf you live with a lame person youwill learn to limprdquo However it is interesting that with thestrengthening of the social network establishment the effectsof online social networks and offline social networks onadolescent sports behavior are not the same When ado-lescents have both an offline social network and an onlinesocial network the offline social network has a positive effecton adolescent sports while the online social network has anegative effect on adolescent sports +is outcome may bebecause at the present stage Chinese adolescents makefriends online mainly through games chats and other ways[21] Moreover adolescents in junior and senior highschools have heavy daily learning tasks and usually surf theInternet for leisure and entertainment +erefore mostadolescentsrsquo online social network relationships are enter-tainment groups Besides during the formation of onlinenetwork relationships most adolescents will use computersor mobile phones for a long time a factor that further re-duces the possibility of adolescent sports [22] Under theinfluence of the characteristics of network groups and thetime used for online communication offline social networkshave a negative influence on adolescent sports behavior

For social networks with different characteristics the in-strumental network emotional network and mixed networkhave positive effects on adolescent sports respectively In otherwords without the influence of others social relations such asthose with counselors teachers family members and friendshave positive effects on adolescent sports behaviors Howeverwhen adolescents have the above three kinds of social relationsat the same time the influence of the instrumental networkthat is counselors or teachers on adolescent sports behavior isnot significant +at is to say the influence of counselors orteachers on adolescent sports behavior is easily affected byother people +is outcome may be because the instrumentalnetwork connectivity is not strong [23] Even for teachers andstudents who live and study very closely there is no similarinfluence like those of friends indicating that although physicalactivity has the interpersonal infectivity of social network itsinfectivity changes based on relative activities and the gendersof friends [24] For example in a given social network lessactive runners tend to affect more active runners while theopposite is not true Both men and women affect men whileonly women affect other women [25]+e reason why it is easyfor physical activity behaviors to spread through adolescentfriendship or a family network is that first of all adolescents aremore willing to experience peer relations during adolescenceand their peer groups are based on common behaviors (sportscomputer games video games diet etc) +ese directly orindirectly affect their physical activities and health [26] Sec-ondly family relationship is the network relationship with thestrongest connection among adolescents Under the influenceof that strong connection an emotional network will have anobvious influence on adolescent sports behavior

Overall different types of social networks have positiveeffects on adolescent sports behavior separately but differentsocial networks are also interacted so their influences onadolescent sports behavior will change by the interactions

Complexity 7

+at is to say the effect of social networks on sports behaviornot only is influenced by the strength and density of socialnetworks but also influenced the interaction of differentsocial networks

332 Mediating Effect of Social Efficacy and the ModeratingEffect of Self-Presentation Social efficacy plays a mediatingrole in the relationship between social networking andadolescent sports behavior and these results can beexplained using the following perspectives

First according to the theoretical model of ldquotwo pathsand three layersrdquo of trust behavior social relationship is asituational factor and its influence on adolescent sportsbehavior is affected by internal cognition [27] In a positiveand extroverted social relationship adolescent sports be-haviors tend to receive more social support which then canenable adolescents to form a more positive internal cogni-tion of themselves and others thereby producing a strongermotivation to undertake sports [28]

Secondly from the perspective of self-efficacy theorysituational condition is one of the important factors that canaffect individual self-efficacy [29] If one is in a familiar andpleasant environment individualrsquos self-efficacy will increaseDoing sports is an important way to relieve adolescentpsychological stress and improve physical health +usadolescents receive positive demonstrations and responsesfrom their social network and their motivation to exercisewill strengthen in order to achieve the recognition of theirsocial relationships

+e moderating effect of self-presentation in the rela-tionship between a social network and adolescent sportsbehavior is not significant in that the positive effect of social

network on adolescent sports behavior is not affected byadolescentsrsquo self-presentation+is outcomemay be becauseon the one hand the level of self-presentation in Chineseadolescentsrsquo sports is low According to the Research Reporton Internet Usersrsquo Information Security released by the ChinaInternet Network Information Center (CNNIC 2018) morethan 73 of adolescentsrsquo self-presentation focuses on en-tertainment and study not so much on sports [30] On theother hand although self-presentation can reduce theloneliness of individuals and show them the enthusiasm ofsports behaviors it also presents certain negative informa-tion about sports such as fatigue and spending time therebyweakening the positive effect of self-presentation [31]

To sum up then this study concludes that social net-working not only directly predicts adolescent sports be-havior but also indirectly affects their actual sports behaviorthrough social efficacy +e implication of this mediatormodel is that adolescent sports behavior can also be gen-erated through interpersonal interaction and that behaviorcan be affected by individual cognition thereby showing therelevance of self-cognition gained by social interaction [32]

4 Conclusions

+e main conclusions of this study are the following

(1) +e social networks of adolescents have obviouspositive effects on their sports behavior Single onlinesocial networks and offline social networks instru-mental networks emotional networks and mixednetworks have obvious positive effects on adolescentsports behavior respectively However under theinfluence of multiple types of social networks offline

Table 4 Estimated results of moderating effect of self-presentation

Variables Adolescent sports behaviorConstant 0383lowast (1654) 0330lowast (1768) 0317lowast (1691)Age minus0028lowast (minus1764) minus0024lowast (minus1936) minus0024lowast (minus1916)Gender 0036 (0673) 0049 (1134) 0046 (1069)Self- presentation 0811lowastlowastlowast (30426) 0332lowastlowastlowast (9093) 0329lowastlowastlowast (8993)Social network mdash 0592lowastlowastlowast (16231) 0589lowastlowastlowast (16109)Social networklowastself-presentation mdash mdash 0014 (0891)Adj-R2 0654 0774 0774Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

Table 3 Estimated results and significance test results of mediating effect of social efficacy

Variables Equation (11)Sports

Equation (12)Social efficacy

Equation (13)Sports

Constant 0316 (1568) 0460lowastlowast (2148) minus0088 (minus1191)Age minus0023lowast (minus1667) minus0032lowastlowast (minus2228) 0006 (1123)Gender 0028 (0599) 0018 (0364) 0012 (0716)Social network 0860lowastlowastlowast (37052) 0839lowastlowastlowast (34099) 0122lowastlowastlowast (7916)Social efficacy mdash mdash 0879lowastlowastlowast (5691)Adj-R2 0737 0706 0956Indirect effect(s) of SN on sport

Intervening variable social contact Effect Boot standard error Boot LLCI Boot ULCI07374 00178 07007 07704

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

8 Complexity

social networking has a negative predictive effect onadolescent sports behavior while the effect of mixednetworks is not so obvious

(2) Social efficacy plays an intermediary role in the re-lationship between social networking and adolescentsports behavior +e moderating effect of self-pre-sentation is not significant

Although this paper found the importance of socialnetworks in adolescents sports behavior and discovered therole of social efficacy there are some limitations to this studywhich need to be addressed in future studies For example allvariables in this study were investigated using a questionnaireit canmeasure variables in static level butmay ignore diversityand complexity of variables furthermore this study foundthat the moderating effect of self-presentation is not signif-icant+is finding may have a theoretical rationale but it mayalso be due to the low degree of self-presentation of thesample In the future that sample needs to be expanded forfurther research and more precise evaluation

Data Availability

+e survey data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

+is study was supported by the MOE (Ministry of Edu-cation in China) Project of Humanities and Social Sciences(Project no 18YJC890012)

References

[1] L P Spear ldquo+e adolescent brain and age-related behavioralmanifestationsrdquo Neuroscience amp Biobehavioral Reviewsvol 24 no 4 pp 417ndash463 2000

[2] L Steinberg and K C Monahan ldquoAge differences in resis-tance to peer influencerdquo Developmental Psychology vol 43no 6 pp 1531ndash1543 2009

[3] S Sarah-Jeanne N R James C B Julie et al ldquoEffect of peersand friends on youth physical activity and motivation to bephysically activerdquo Journal of Pediatric Psychology vol 34no 2 pp 217ndash225 2009

[4] M Jennifer D L H Kayla L M Barnett et al ldquoFriendshipnetwork characteristics are associated with physical activityand sedentary behavior in early adolescencerdquo PLoS ONEvol 10 no 12 pp 1ndash15 2015

[5] M J Edwards R Jago J S Simon et al ldquo+e influence offriends and siblings on the physical activity and screenviewing behaviours of children aged 5-6 years a qualitativeanalysis of parent interviewsrdquo BMJ Open vol 5 no 5 ArticleID e006593 2015

[6] K D L Haye G Robins PMohr et al ldquoHow physical activityshapes and is shaped by adolescent friendshipsrdquo SocialScience amp Medicine vol 73 no 5 pp 719ndash728 2011

[7] Z Li F Xiong X Wang et al ldquoTopological influence-awarerecommendation on social networksrdquo Complexity vol 2019Article ID 6325654 12 pages 2019

[8] B Yu and X Hu ldquoToward training and assessing reproducibledata analysis in data science educationrdquo Data Intelligencevol 1 no 4 pp 333ndash344 2019

[9] A Bandura ldquoSelf-efficacy mechanism in human agencyrdquoAmerican Psychologist vol 37 no 2 pp 122ndash147 1982

[10] A Bandura Self-efficacy He Exercise of Control WorthPublishers New Nork NY USA 1997

[11] C Fan and A S Mak ldquoMeasuring social self-efficacy in aculturally diverse student populationrdquo Social Behavior andPersonality An International Journal vol 26 no 2 pp 131ndash144 1998

[12] M Wei D W Russell and R A Zakalik ldquoAdult attachmentsocial self-efficacy self-disclosure loneliness and subsequentdepression for freshman college students a longitudinal studyrdquoJournal of Counseling Psychology vol 52 no 4 pp 602ndash614 2005

[13] F Xiong W Shen H Chen et al ldquoExploiting implicit in-fluence from information propagation for social recom-mendationrdquo IEEE Transactions on Cybernetics 2019

[14] H Y Mao H C Hsu and S D Lee ldquoGender differences inrelated influential factors of regular exercise behavior amongpeople in Taiwan in 2007 a cross-sectional studyrdquo PLoS onevol 15 no 1 pp 1ndash22 2020

[15] S Park J Y Kang and L A Chadiha ldquoSocial network typeshealth and health-Care use among South Korean olderadultsrdquo Research on Aging vol 40 no 2 pp 131ndash154 2018

[16] E J Jeong and D H Kim ldquoSocial activities self-efficacy gameattitudes and game addictionrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 4 pp 213ndash221 2011

[17] J Kim and J-E R Lee ldquo+e Facebook paths to happinesseffects of the number of Facebook friends and self-presen-tation on subjective well-beingrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 6 pp 359ndash364 2011

[18] Z Wen and B Ye ldquoAnalyses of mediating effects the de-velopment of methods and modelsrdquo Advances in Psycho-logical Science vol 22 no 5 pp 731ndash745 2014

[19] L G Smith L Banting R Eime et al ldquo+e association be-tween social support and physical activity in older adultsasystematic reviewrdquo International Journal of Behavioral Nu-trition amp Physical Activity vol 14 no 1 p 56 2017

[20] K P Smith and N A Christakis ldquoSocial networks andhealthrdquo Annual Review of Sociology vol 34 no 1pp 405ndash429 2008

[21] S Aral and C Nicolaides ldquoExercise contagion in a globalsocial networkrdquo Nature Communications vol 8 Article ID14753 2017

[22] S Ji S Pan E Cambria et al ldquoA survey on knowledge graphsrepresentation acquisition and applicationsrdquo 2020 httparxivorgabs200200388

[23] T W Valente K Fujimoto C-P Chou and D Spruijt-MetzldquoAdolescent affiliations and adiposity a social networkanalysis of friendships and obesityrdquo Journal of AdolescentHealth vol 45 no 2 pp 202ndash204 2009

[24] P Manasatchakun P Chotiga J Hochwalder A RoxbergM Sandborgh and M Asp ldquoFactors associated with healthyaging among older persons in northeastern+ailandrdquo Journalof Cross-Cultural Gerontology vol 31 no 4 pp 369ndash3842016

[25] L Adam and P S Damon ldquo+e intersection of sport man-agement and sociology of sport research a social networkperspectiverdquo Sport Management Review vol 15 no 2pp 244ndash256 2012

Complexity 9

At the same time to deeply analyze the influence ofdifferent social networks on adolescents sports this studydivides social networks into online and offline social plat-forms and emotional instrumental and mixed social net-works Of these teachers and counselors belong to theinstrumental network family members and lovers belong tothe emotional network and roommates and classmatesbelong to a mixed network [11] +rough an analysis ofsocial networks social efficacy and self-presentation thispaper offers both a theoretical basis and practical guidancefor promoting the establishment of social networks andfacilitating adolescent sports behavior

2 Materials and Methodology

21 Subjects +e subjects were 630 adolescents from threejunior middle schools and three senior high schools inShaanxi Henan and Shandong Provinces A total of 630questionnaires were distributed and of these 568 validquestionnaires were collected providing an effective re-covery rate of 901 +ere were 296 boys (5211) and 272girls (4789) 376 senior students and 192 junior students+e average age of the subjects was 1452 years (SD 168with the age range being 12ndash18)

22 Measurements An adolescent sports behavior scale isbased on the questionnaire by Mao et al [14] which in-cluded 6 items (eg I usually take exercise with friends Ihave the habit of exercising) and the questionnaire wasscored using 5 answers (1 ldquototally inconsistentrdquo to 5ldquocompletely consistentrdquo) In this study the confirmatoryfactor analysis of the questionnaire fits well χ2df 284RMSEA 005 NFI 0995 GFI 0998 and CFI 0997+e internal concordance coefficient α was 0687

+e social network scale uses a questionnaire designedby Park et al [15] which contains 14 items and is dividedinto either two dimensions as (1) an online social network (6items eg Many of my friends in real life love sports) and (2)an offline social network (4 items eg +e number ofcommunities Irsquove join online) or three dimensions as (1) aninstrumental social network (4 items eg I have a closerelationship with my class teacher) (2) an emotional net-work (4 items eg I have a close relationship with myrelatives of the same age) and (3) a mixed network (4 itemseg I have a close relationship with my friends in real life)

Some of the items in the questionnaire on different di-mensions were the same For example ldquoI have a close rela-tionship with my friends in real liferdquo is a measurement item ofboth an offline social network and a mixed network +equestionnaire is scored using 5 points (1 ldquototally inconsistentrdquoto 5 ldquocompletely consistentrdquo) In the current study theconfirmatory factor analysis of the questionnaire fits wellχ2df 284 RMSEA 006 NFI 0998 GFI 0999 andCFI 0999+e internal concordance coefficient αwas 0654+e coefficients α of each dimension were 0940 (online socialnetwork) 0936 (offline social network) 0898 (instrumentalsocial network) 0766 (emotional social network) and 0847(mixed social network)

+e social efficacy scale used a questionnaire compiledby Jeong and Kim [16] +ere were 6 items on this ques-tionnaire in the study (eg online communication I caneasily become friends with other people I can easily talk tounfamiliar people) +e questionnaire was scored using 5points (1 ldquototally inconsistent ldquoto 5rdquo completely consistentrdquo)In this study the confirmatory factor analysis of the ques-tionnaire fits well χ2df 124 RMSEA 002 NFI 0998GFI 0999 and CFI 0999 +e internal concordancecoefficient α was 0634

+e self-presentation scale used a questionnaire com-piled by Kim and Lee [17] It included 4 items (eg I will postphotos that show the real me I donrsquot mind sharing some badthings that happened to me on the Internet) +e ques-tionnaire was scored based on 5 points (1 ldquototally incon-sistentrdquo to 5 ldquocompletely consistentrdquo) In this study theconfirmatory factor analysis of the questionnaire fits wellχ2df 224 RMSEA 002 NFI 0998 GFI 0999 andCFI 0999 +e internal concordance coefficient α was0820

23 Programs and Data Processing In this current studyafter obtaining the consent of the school leaders teachersand students the questionnaire practice was explained byhighly trained surveyors in accordance with standardizedinstruction and all questionnaires were collected immedi-ately SPSS 220 software was used to analyze the data andPROCESS V30 was used to test the mediating effect

First we used the factor analysis method to reduce thedimensionality of variables +e mathematical model was

x1 a11f1 + a12f2 + a13f3 + middot middot middot + a1kfk + ε1

x2 a21f1 + a22f2 + a23f3 + middot middot middot + a2kfk + ε2

x3 a31f1 + a32f2 + a33f3 + middot middot middot + a3kfk + ε3

xp ap1f1 + ap2f2 + ap3f3 + middot middot middot + apkfk + εp

⎧⎪⎪⎪⎪⎪⎪⎪⎪⎨

⎪⎪⎪⎪⎪⎪⎪⎪⎩

(1)

For equation (1) xi is the standardized variable fi is thefactor variable ε is the special factor aij is the factor loadkltp A matrix was used to simplify equation (1) as follows

X AF + ε (2)

where F is the factor variable matrix and A is a factor loadmatrix and satisfies cov(F ε) 0 D(F) Im D(ε) δi

+e factor load aij represents the degree of correlationbetween Xi and Fi and the square sum of the elements in thei-th row of the factor load matrix where A is h2

i 1113936kj1 a2

ij which represents the explanatory power of all the factorvariables for the total variance of Xi and Sj 1113936

p

i1 a2ij is the

variance contribution of the variable FjAfter determining the factor variables the factor load

matrix needed to be estimated If we set the eigenvalues ofthe sample covariance matrix and the corresponding stan-dard orthogonalized eigenvectors as λ1 ge λ2 ge middot middot middot ge λp ge 0e1 e2 ep then the covariance matrix can be decomposedinto

Complexity 3

1113944 U

λ1 0

λ2

0 λp

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝

⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠Uprime 1113944

p

i1λieieiprime (3)

When the last several eigenvalues are small the co-variance matrix can be approximately decomposed into

1113944 asymp

λ11113969

e1

λm

1113969

em1113874 1113875

λ1

1113968e1prime

λm

1113968emprime

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦+

σ21σ22

σ2p

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝

⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

AAprime + 1113944ε

(4)

where A is the factor covariance matrix Since the factorloading matrix is not unique factor rotation was executed tomake the meaning of the common factor clearer Consid-ering the orthogonal rotation of two factors the orthogonalmatrix then became

Q cos ϕ minussinϕ

sinϕ cosϕ11138891113888 (5)

and set

B AQ bij11138731113872 i 1 2 p j 1 2 (6)

where B is the rotation factor load matrix At this time werequired the variance of the two columnsrsquo data in B to be aslarge as possible which also meant that the relative varianceVi should also be as large as possible

Vj 1p

1113944

p

i1

b2ij

h2i

11138891113888

2

minus 1113944

p

i1

b2ij

h2i

11138891113888

2

j 1 2 (7)

Making dVdϕ 0 then ϕ should satisfy the followingequation

tan 4ϕ D0 minus 2A0B0( 1113857p( 1113857

C0 minus A20 minus B2

0( 1113857p( 1113857 (8)

where

A0 1113936p

i1ui B0 1113936

p

i1vi

C0 1113936p

i1u2

i minus v2i( 1113857 D0 2 1113936p

i1uivi

ui a2

i1hi

1113874 11138752

minusa2

i2hi

1113874 11138752 vi

2ai1ai2

h2i

⎧⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎨

⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎩

(9)

Using SPSS for the factor analysis taking adolescentsports and social network as examples the scree plots wereobtained It can be seen from Figure 1 that the adolescent

sports scale can extract one factor while the social networkscale can extract two factors

3 Results and Discussion

31 Descriptive Analysis +e mean standard deviation andthe correlation coefficient of each variable are shown inTable 1 +ese results show that there are obvious correla-tions between adolescent sports behavior and different socialnetworks and also obvious correlations between social ef-ficacy self-presentation social networking and sports Inaddition social networks especially offline social networkshave significant age differences Different types of socialnetworks and self-presentations also have gender differ-ences In order to explore the independent effects of socialnetwork on adolescent sports age and gender were used hereas the control variables

Furthermore we use histogram to basically analyze thesample From Figure 2(a) it can be seen that the sample isbalanced and distributed between male and female indi-cating the sample can reflect reality well In different gendergroup (Figure 2(b)) the number of adolescents who are 13-14 years old is the most while the number of 12 years oldadolescents is the least

From Figure 2(c) during the group of 12 years oldadolescents the scores of five social networks are all highindicating that for 12-year-olds their social network typedensity and intensity are stronger For 13-14 years oldadolescents all types of social networks are not high es-pecially online social network Its average value is 12 whichmeans the online social network is not strong +is maybecause the academic pressure for adolescents in the stage ishigher than other groups For 15-16 years old adolescentsmixed social network and emotional social network arestronger than other social networks and for 17-18 years oldadolescents online social network mixed social networkand emotional social network are stronger than other twosocial networks From Figure 2(d) no matter how old ad-olescents are their sports activities are frequent For ex-ample for 13-14 years old adolescents most adolescents dosports more than once a week only 5 of adolescents dosports once a week or even less than once a week FromFigure 2 we can clearly know the distribution of sample thesocial networks of different group and the sports behavior ofdifferent groups it helps to lay the foundation for the fol-lowing empirical research

32 Results

321 Influence of Social Networking on Adolescent SportsBehavior Econometric models are constructed to test therelationships between different social networks and ado-lescent sports behavior

Sporti βXi + PQ + e1 (10)

Among them Sport is adolescent sports behavior X isan adolescentsrsquo social network that includes an onlineofflinesocial network and an instrumentalemotionalmixed social

4 Complexity

network Q is the control variable matrix and e1 is theregression residual +e empirical results are shown inTable 2

+e results in Table 2 indicate that adolescentsrsquo socialnetworks have a significant positive effect on their sportsbehavior (β 0860 plt 0001) +e wider the homoge-neity of these adolescentsrsquo social relationship is the morefrequent their sports behavior becomes Online socialnetworks and offline social networks have positive effectson adolescent sports behavior respectively Howeverwhen adolescents have both an offline social network andan online social network the offline social network has apositive effect on adolescent sports (β 1381 plt 0001)while the online social network has a negative effect onadolescent sports (β 0517 plt 0001) An instrumentalnetwork emotional network and mixed networkhave positive effects on adolescent sports respectivelywith coefficients of 0735 0964 0880 and are significantat a 1 level However when adolescents have the threekinds of above social relations at the same time theemotional network and the mixed network still havepositive effects on adolescent sports but the influence ofthe instrumental network on adolescent sports then be-comes insignificant

322 Mediating Effect of Social Efficacy According to Wenand Ye [18] it is necessary to test the parameters of threeregression equations to verify the mediating effect

Sporti cSNi + PQ + e2 (11)

SEi αSNi + PQ + e3 (12)

Sporti cprimeSPi + bSEi + PQ + e4 (13)

Among these SN is social network SE is social efficacyQ is the control variable matrix and e2 sim e4 are the re-gression residuals Further c is the total effect of the in-dependent variable (social network) on the dependentvariable (adolescent sports behavior) a is the effect of socialnetwork on the intervening variable (social efficacy) b is theeffect of social efficacy on adolescent sports behavior aftercontrolling for the influence of social network cprime is the directeffect of social network on adolescent sports behavior aftercontrolling for the influence of social efficacy +e mediatingeffect is tested in five steps first test the coefficient c ofequation (11) if c is significant the intermediary effect issignificant otherwise there is a masking effect But whetherit is significant or not follow-up tests are carried out

1 2 3 4Number

Scree plot

000

050

100

150

200

250

300

350

Eige

nval

ues

(a)

1 2 3 4 5 6 7 8 9 10 11Number

Scree plot

000100200300400500600700800

Eige

nval

ues

(b)

Figure 1 Scree plot of adolescent sports (a) and social network (b)

Table 1 Mean standard deviation and correlation coefficient of variables

M SD 1 2 3 4 5 6 7 8 9 10 111 Age 1453 169 12 Gender 053 050 minus0018 13 Sports 288 093 minus0042 minus0076 14 Social network 208 087 minus0004lowast minus0105lowast 0858lowast 15 Online social network 209 089 0015lowast minus0084 0744lowastlowast 0967lowastlowast 16 Offline social network 236 064 minus0015 minus0113lowast 0903lowastlowast 0988lowastlowast 0916lowastlowast 17 Mixes social network 230 071 minus0018 minus0126lowastlowast 0876lowastlowast 0945lowast 0863lowastlowast 0966lowastlowast 18 Instrumental socialnetwork 234 066 0000 minus0107lowast 0734lowastlowast 0944lowastlowast 0923lowastlowast 0924lowastlowast 0823lowastlowast 1

9 Emotional socialnetwork 1453 169 minus0033 minus0066 0965lowastlowast 0844lowastlowast 0724lowastlowast 0891lowastlowast 0850lowastlowast 0711lowastlowast 1

10 Social efficacy 241 071 minus0058 minus0078 0980lowastlowast 0838lowastlowast 0725lowastlowast 0884lowastlowast 0880lowastlowast 0712lowastlowast 0916lowastlowast 111 Self-presentation 243 096 0006 minus0116lowastlowast 0808lowastlowast 0810lowastlowast 0710lowastlowast 0848lowastlowast 0895lowastlowast 0662lowastlowast 0816lowastlowast 0780lowastlowast 1Note N 568 gender is a dummy variable female 0 and male 1 lowastplt 005 lowastlowastplt 001

Complexity 5

Second test the coefficient a of equation (12) and the co-efficient b of equation (13) in turn if both are significantthen the indirect effect is significant go to the fourth step ifat least one is not significant go to the third step +ird useBootstrap method to directly check H0 ab 0 If it issignificant then the indirect effect is significant go to thefourth step otherwise the indirect effect is not significantstop the analysis Forth examining the coefficient cprime ofequation (13) if it is not significant the direct effect is not

significant indicating that there is only an intermediaryeffect If it is significant that is the direct effect is significantgo to the fifth step Fifth compare the signs of ab and cprime Ifthey are the same it is partly intermediary effect If the sign isdifferent it is a masking effect If a b and c are all significantthe mediation effect is significant otherwise there are othereffects For example if c is not significant there is a maskingeffect if cprime is significant there also exist direct effects Besidesthe above five steps a nonparametric percentile Bootstrap

Below 12 13-14 15-16 17-18

Statistics of gender in different ages

MaleFemale

010203040506070

()

(a)

Male Female

Statistics of ages in different genders

Below 1213-14

15-1617-18

05

101520253035404550

()

(b)

Online social networkOffline social networkMixed social network

Instrumental social networkEmotional social network

Below 12 13-14 15-16 17-18

Statistics of social networks in different ages

0

1

2

3

4

5

6

(c)

Once a weekOne to three times a weekFour to five times a weekMore than five times a week

Below 12 13-14 15-16 17-18

Statistics of sport behaviors in different ages

05

10152025303540

()

(d)

Figure 2 Statistic of adolescent social networks and sports behavior

Table 2 Estimated results of the influence of social network on adolescent sports behavior

Variables Adolescent sports behaviorConstant 0316 (1568) 0471lowast (1794) 0210 (1240) 0069 (0465) 0357 (1336) 0100 (0966) 0179 (0945) 0102 (1071)Age minus0023lowast (minus1667) minus0031lowast (-1767) minus0016 (minus1427) minus0007 (minus0737) minus0025 (minus1366) minus0006 (minus0859) minus0015 (minus1157) minus0007 (minus1113)Gender 0028 (0599) minus0028 (minus0458) 0053 (1356) 0074lowastlowast (2154) 0004 (0068) minus0024 (minus1004) 0070 (1596) 0004 (0203)Socialnetwork 0860lowastlowastlowast (37052) mdash mdash mdash mdash mdash mdash mdash

Online socialnetwork mdash 0744lowastlowastlowast (24673) mdash minus0517lowastlowastlowast (minus12194) mdash mdash mdash mdash

Offline socialnetwork mdash mdash 0905lowastlowastlowast (45545) 1381lowastlowastlowast (32455) mdash mdash mdash mdash

Instrumentalnetwork mdash mdash mdash mdash 0735lowastlowastlowast (23892) mdash mdash 0016 (0818)

Emotionalnetwork mdash mdash mdash mdash mdash 0964lowastlowastlowast (81163) mdash 0793lowastlowastlowast (38244)

Mixednetwork mdash mdash mdash mdash mdash mdash 0880lowastlowastlowast (40217) 0189lowastlowastlowast (7352)

Adj-R2 0737 0554 0816 0858 0538 0931 0767 0942

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

6 Complexity

method was used to test the mediating effect If the 95confidence interval does not contain a zero then the me-diating effect becomes significant +e empirical regressionresults and significance test results for the mediating effectare shown in Table 3

+e results in Table 3 show that R2 in the three equationswas bigger than 07 and the goodness of fit for the model wasgood In equation (11) social network has a significantpositive effect on adolescent sports behaviorc 0860 plt 0001 In equation (12) social network has asignificant effect on social efficacy c 0839 plt 0001 Inequation (13) social network and social efficacy have asignificant effect on adolescent sports behaviorcprime 0122 b 0879 plt 0001 According to the mediatingeffect test procedure of Wen and Ye [18] the mediatingeffect of social efficacy is significant because a b c and cprime areall significant Further the Bootstrap method was used totest the mediating effect of social efficacy 5000 Bootstrapsamples were randomly selected from the original sample forindirect effect estimation Table 3 shows the estimated valueof the indirect effect of social efficacy to be 07374 and 95confidence intervals were [07007 07704] excluding zero+us the mediating effect of social efficacy becamesignificant

323 Moderating Effect of Self-Presentation As for themoderating effect of self-presentation the followingeconometric model was constructed

Sporti αSNi + βSPi + cSNi lowast SPi + PQ + e5 (14)

where SP is the self-presentation +e stepwise regressionmethod was used to estimate the moderating effect of self-presentation +e results are shown in Table 4

+e results of the study show that although both socialnetworking and self-presentation have significant positiveeffects on adolescent sports behavior the interaction termsof social network and self-presentation do not have anysignificant influence on adolescent sports behavior therebyindicating that the moderating effect of self-presentation isnot significant

In addition this study found that age has a negativepredictive effect on adolescent sports behavior that is theolder the adolescents are the less likely they are to exerciseGender had no significant effect on the prediction of ado-lescent sports behavior

33 Discussion

331 Influence of Social Networks on Adolescent SportsBehavior +e results of this study indicate that socialnetworks have significant influences on adolescent sportsbehavior demonstrating that adolescents with larger anddenser social networks have more frequent sports behaviorPeople are mutually linked to each other and so are theirsports behaviors +us the existence of social networksmeans that personsrsquo movements are interdependent [19]Sports behavior can be spread from person to person and

there is also a phenomenon called ldquopeer effectsrdquo [20] whichis just as the proverb says ldquoIf you live with a lame person youwill learn to limprdquo However it is interesting that with thestrengthening of the social network establishment the effectsof online social networks and offline social networks onadolescent sports behavior are not the same When ado-lescents have both an offline social network and an onlinesocial network the offline social network has a positive effecton adolescent sports while the online social network has anegative effect on adolescent sports +is outcome may bebecause at the present stage Chinese adolescents makefriends online mainly through games chats and other ways[21] Moreover adolescents in junior and senior highschools have heavy daily learning tasks and usually surf theInternet for leisure and entertainment +erefore mostadolescentsrsquo online social network relationships are enter-tainment groups Besides during the formation of onlinenetwork relationships most adolescents will use computersor mobile phones for a long time a factor that further re-duces the possibility of adolescent sports [22] Under theinfluence of the characteristics of network groups and thetime used for online communication offline social networkshave a negative influence on adolescent sports behavior

For social networks with different characteristics the in-strumental network emotional network and mixed networkhave positive effects on adolescent sports respectively In otherwords without the influence of others social relations such asthose with counselors teachers family members and friendshave positive effects on adolescent sports behaviors Howeverwhen adolescents have the above three kinds of social relationsat the same time the influence of the instrumental networkthat is counselors or teachers on adolescent sports behavior isnot significant +at is to say the influence of counselors orteachers on adolescent sports behavior is easily affected byother people +is outcome may be because the instrumentalnetwork connectivity is not strong [23] Even for teachers andstudents who live and study very closely there is no similarinfluence like those of friends indicating that although physicalactivity has the interpersonal infectivity of social network itsinfectivity changes based on relative activities and the gendersof friends [24] For example in a given social network lessactive runners tend to affect more active runners while theopposite is not true Both men and women affect men whileonly women affect other women [25]+e reason why it is easyfor physical activity behaviors to spread through adolescentfriendship or a family network is that first of all adolescents aremore willing to experience peer relations during adolescenceand their peer groups are based on common behaviors (sportscomputer games video games diet etc) +ese directly orindirectly affect their physical activities and health [26] Sec-ondly family relationship is the network relationship with thestrongest connection among adolescents Under the influenceof that strong connection an emotional network will have anobvious influence on adolescent sports behavior

Overall different types of social networks have positiveeffects on adolescent sports behavior separately but differentsocial networks are also interacted so their influences onadolescent sports behavior will change by the interactions

Complexity 7

+at is to say the effect of social networks on sports behaviornot only is influenced by the strength and density of socialnetworks but also influenced the interaction of differentsocial networks

332 Mediating Effect of Social Efficacy and the ModeratingEffect of Self-Presentation Social efficacy plays a mediatingrole in the relationship between social networking andadolescent sports behavior and these results can beexplained using the following perspectives

First according to the theoretical model of ldquotwo pathsand three layersrdquo of trust behavior social relationship is asituational factor and its influence on adolescent sportsbehavior is affected by internal cognition [27] In a positiveand extroverted social relationship adolescent sports be-haviors tend to receive more social support which then canenable adolescents to form a more positive internal cogni-tion of themselves and others thereby producing a strongermotivation to undertake sports [28]

Secondly from the perspective of self-efficacy theorysituational condition is one of the important factors that canaffect individual self-efficacy [29] If one is in a familiar andpleasant environment individualrsquos self-efficacy will increaseDoing sports is an important way to relieve adolescentpsychological stress and improve physical health +usadolescents receive positive demonstrations and responsesfrom their social network and their motivation to exercisewill strengthen in order to achieve the recognition of theirsocial relationships

+e moderating effect of self-presentation in the rela-tionship between a social network and adolescent sportsbehavior is not significant in that the positive effect of social

network on adolescent sports behavior is not affected byadolescentsrsquo self-presentation+is outcomemay be becauseon the one hand the level of self-presentation in Chineseadolescentsrsquo sports is low According to the Research Reporton Internet Usersrsquo Information Security released by the ChinaInternet Network Information Center (CNNIC 2018) morethan 73 of adolescentsrsquo self-presentation focuses on en-tertainment and study not so much on sports [30] On theother hand although self-presentation can reduce theloneliness of individuals and show them the enthusiasm ofsports behaviors it also presents certain negative informa-tion about sports such as fatigue and spending time therebyweakening the positive effect of self-presentation [31]

To sum up then this study concludes that social net-working not only directly predicts adolescent sports be-havior but also indirectly affects their actual sports behaviorthrough social efficacy +e implication of this mediatormodel is that adolescent sports behavior can also be gen-erated through interpersonal interaction and that behaviorcan be affected by individual cognition thereby showing therelevance of self-cognition gained by social interaction [32]

4 Conclusions

+e main conclusions of this study are the following

(1) +e social networks of adolescents have obviouspositive effects on their sports behavior Single onlinesocial networks and offline social networks instru-mental networks emotional networks and mixednetworks have obvious positive effects on adolescentsports behavior respectively However under theinfluence of multiple types of social networks offline

Table 4 Estimated results of moderating effect of self-presentation

Variables Adolescent sports behaviorConstant 0383lowast (1654) 0330lowast (1768) 0317lowast (1691)Age minus0028lowast (minus1764) minus0024lowast (minus1936) minus0024lowast (minus1916)Gender 0036 (0673) 0049 (1134) 0046 (1069)Self- presentation 0811lowastlowastlowast (30426) 0332lowastlowastlowast (9093) 0329lowastlowastlowast (8993)Social network mdash 0592lowastlowastlowast (16231) 0589lowastlowastlowast (16109)Social networklowastself-presentation mdash mdash 0014 (0891)Adj-R2 0654 0774 0774Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

Table 3 Estimated results and significance test results of mediating effect of social efficacy

Variables Equation (11)Sports

Equation (12)Social efficacy

Equation (13)Sports

Constant 0316 (1568) 0460lowastlowast (2148) minus0088 (minus1191)Age minus0023lowast (minus1667) minus0032lowastlowast (minus2228) 0006 (1123)Gender 0028 (0599) 0018 (0364) 0012 (0716)Social network 0860lowastlowastlowast (37052) 0839lowastlowastlowast (34099) 0122lowastlowastlowast (7916)Social efficacy mdash mdash 0879lowastlowastlowast (5691)Adj-R2 0737 0706 0956Indirect effect(s) of SN on sport

Intervening variable social contact Effect Boot standard error Boot LLCI Boot ULCI07374 00178 07007 07704

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

8 Complexity

social networking has a negative predictive effect onadolescent sports behavior while the effect of mixednetworks is not so obvious

(2) Social efficacy plays an intermediary role in the re-lationship between social networking and adolescentsports behavior +e moderating effect of self-pre-sentation is not significant

Although this paper found the importance of socialnetworks in adolescents sports behavior and discovered therole of social efficacy there are some limitations to this studywhich need to be addressed in future studies For example allvariables in this study were investigated using a questionnaireit canmeasure variables in static level butmay ignore diversityand complexity of variables furthermore this study foundthat the moderating effect of self-presentation is not signif-icant+is finding may have a theoretical rationale but it mayalso be due to the low degree of self-presentation of thesample In the future that sample needs to be expanded forfurther research and more precise evaluation

Data Availability

+e survey data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

+is study was supported by the MOE (Ministry of Edu-cation in China) Project of Humanities and Social Sciences(Project no 18YJC890012)

References

[1] L P Spear ldquo+e adolescent brain and age-related behavioralmanifestationsrdquo Neuroscience amp Biobehavioral Reviewsvol 24 no 4 pp 417ndash463 2000

[2] L Steinberg and K C Monahan ldquoAge differences in resis-tance to peer influencerdquo Developmental Psychology vol 43no 6 pp 1531ndash1543 2009

[3] S Sarah-Jeanne N R James C B Julie et al ldquoEffect of peersand friends on youth physical activity and motivation to bephysically activerdquo Journal of Pediatric Psychology vol 34no 2 pp 217ndash225 2009

[4] M Jennifer D L H Kayla L M Barnett et al ldquoFriendshipnetwork characteristics are associated with physical activityand sedentary behavior in early adolescencerdquo PLoS ONEvol 10 no 12 pp 1ndash15 2015

[5] M J Edwards R Jago J S Simon et al ldquo+e influence offriends and siblings on the physical activity and screenviewing behaviours of children aged 5-6 years a qualitativeanalysis of parent interviewsrdquo BMJ Open vol 5 no 5 ArticleID e006593 2015

[6] K D L Haye G Robins PMohr et al ldquoHow physical activityshapes and is shaped by adolescent friendshipsrdquo SocialScience amp Medicine vol 73 no 5 pp 719ndash728 2011

[7] Z Li F Xiong X Wang et al ldquoTopological influence-awarerecommendation on social networksrdquo Complexity vol 2019Article ID 6325654 12 pages 2019

[8] B Yu and X Hu ldquoToward training and assessing reproducibledata analysis in data science educationrdquo Data Intelligencevol 1 no 4 pp 333ndash344 2019

[9] A Bandura ldquoSelf-efficacy mechanism in human agencyrdquoAmerican Psychologist vol 37 no 2 pp 122ndash147 1982

[10] A Bandura Self-efficacy He Exercise of Control WorthPublishers New Nork NY USA 1997

[11] C Fan and A S Mak ldquoMeasuring social self-efficacy in aculturally diverse student populationrdquo Social Behavior andPersonality An International Journal vol 26 no 2 pp 131ndash144 1998

[12] M Wei D W Russell and R A Zakalik ldquoAdult attachmentsocial self-efficacy self-disclosure loneliness and subsequentdepression for freshman college students a longitudinal studyrdquoJournal of Counseling Psychology vol 52 no 4 pp 602ndash614 2005

[13] F Xiong W Shen H Chen et al ldquoExploiting implicit in-fluence from information propagation for social recom-mendationrdquo IEEE Transactions on Cybernetics 2019

[14] H Y Mao H C Hsu and S D Lee ldquoGender differences inrelated influential factors of regular exercise behavior amongpeople in Taiwan in 2007 a cross-sectional studyrdquo PLoS onevol 15 no 1 pp 1ndash22 2020

[15] S Park J Y Kang and L A Chadiha ldquoSocial network typeshealth and health-Care use among South Korean olderadultsrdquo Research on Aging vol 40 no 2 pp 131ndash154 2018

[16] E J Jeong and D H Kim ldquoSocial activities self-efficacy gameattitudes and game addictionrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 4 pp 213ndash221 2011

[17] J Kim and J-E R Lee ldquo+e Facebook paths to happinesseffects of the number of Facebook friends and self-presen-tation on subjective well-beingrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 6 pp 359ndash364 2011

[18] Z Wen and B Ye ldquoAnalyses of mediating effects the de-velopment of methods and modelsrdquo Advances in Psycho-logical Science vol 22 no 5 pp 731ndash745 2014

[19] L G Smith L Banting R Eime et al ldquo+e association be-tween social support and physical activity in older adultsasystematic reviewrdquo International Journal of Behavioral Nu-trition amp Physical Activity vol 14 no 1 p 56 2017

[20] K P Smith and N A Christakis ldquoSocial networks andhealthrdquo Annual Review of Sociology vol 34 no 1pp 405ndash429 2008

[21] S Aral and C Nicolaides ldquoExercise contagion in a globalsocial networkrdquo Nature Communications vol 8 Article ID14753 2017

[22] S Ji S Pan E Cambria et al ldquoA survey on knowledge graphsrepresentation acquisition and applicationsrdquo 2020 httparxivorgabs200200388

[23] T W Valente K Fujimoto C-P Chou and D Spruijt-MetzldquoAdolescent affiliations and adiposity a social networkanalysis of friendships and obesityrdquo Journal of AdolescentHealth vol 45 no 2 pp 202ndash204 2009

[24] P Manasatchakun P Chotiga J Hochwalder A RoxbergM Sandborgh and M Asp ldquoFactors associated with healthyaging among older persons in northeastern+ailandrdquo Journalof Cross-Cultural Gerontology vol 31 no 4 pp 369ndash3842016

[25] L Adam and P S Damon ldquo+e intersection of sport man-agement and sociology of sport research a social networkperspectiverdquo Sport Management Review vol 15 no 2pp 244ndash256 2012

Complexity 9

1113944 U

λ1 0

λ2

0 λp

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝

⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠Uprime 1113944

p

i1λieieiprime (3)

When the last several eigenvalues are small the co-variance matrix can be approximately decomposed into

1113944 asymp

λ11113969

e1

λm

1113969

em1113874 1113875

λ1

1113968e1prime

λm

1113968emprime

⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣

⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦+

σ21σ22

σ2p

⎛⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎜⎝

⎞⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎟⎠

AAprime + 1113944ε

(4)

where A is the factor covariance matrix Since the factorloading matrix is not unique factor rotation was executed tomake the meaning of the common factor clearer Consid-ering the orthogonal rotation of two factors the orthogonalmatrix then became

Q cos ϕ minussinϕ

sinϕ cosϕ11138891113888 (5)

and set

B AQ bij11138731113872 i 1 2 p j 1 2 (6)

where B is the rotation factor load matrix At this time werequired the variance of the two columnsrsquo data in B to be aslarge as possible which also meant that the relative varianceVi should also be as large as possible

Vj 1p

1113944

p

i1

b2ij

h2i

11138891113888

2

minus 1113944

p

i1

b2ij

h2i

11138891113888

2

j 1 2 (7)

Making dVdϕ 0 then ϕ should satisfy the followingequation

tan 4ϕ D0 minus 2A0B0( 1113857p( 1113857

C0 minus A20 minus B2

0( 1113857p( 1113857 (8)

where

A0 1113936p

i1ui B0 1113936

p

i1vi

C0 1113936p

i1u2

i minus v2i( 1113857 D0 2 1113936p

i1uivi

ui a2

i1hi

1113874 11138752

minusa2

i2hi

1113874 11138752 vi

2ai1ai2

h2i

⎧⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎨

⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎩

(9)

Using SPSS for the factor analysis taking adolescentsports and social network as examples the scree plots wereobtained It can be seen from Figure 1 that the adolescent

sports scale can extract one factor while the social networkscale can extract two factors

3 Results and Discussion

31 Descriptive Analysis +e mean standard deviation andthe correlation coefficient of each variable are shown inTable 1 +ese results show that there are obvious correla-tions between adolescent sports behavior and different socialnetworks and also obvious correlations between social ef-ficacy self-presentation social networking and sports Inaddition social networks especially offline social networkshave significant age differences Different types of socialnetworks and self-presentations also have gender differ-ences In order to explore the independent effects of socialnetwork on adolescent sports age and gender were used hereas the control variables

Furthermore we use histogram to basically analyze thesample From Figure 2(a) it can be seen that the sample isbalanced and distributed between male and female indi-cating the sample can reflect reality well In different gendergroup (Figure 2(b)) the number of adolescents who are 13-14 years old is the most while the number of 12 years oldadolescents is the least

From Figure 2(c) during the group of 12 years oldadolescents the scores of five social networks are all highindicating that for 12-year-olds their social network typedensity and intensity are stronger For 13-14 years oldadolescents all types of social networks are not high es-pecially online social network Its average value is 12 whichmeans the online social network is not strong +is maybecause the academic pressure for adolescents in the stage ishigher than other groups For 15-16 years old adolescentsmixed social network and emotional social network arestronger than other social networks and for 17-18 years oldadolescents online social network mixed social networkand emotional social network are stronger than other twosocial networks From Figure 2(d) no matter how old ad-olescents are their sports activities are frequent For ex-ample for 13-14 years old adolescents most adolescents dosports more than once a week only 5 of adolescents dosports once a week or even less than once a week FromFigure 2 we can clearly know the distribution of sample thesocial networks of different group and the sports behavior ofdifferent groups it helps to lay the foundation for the fol-lowing empirical research

32 Results

321 Influence of Social Networking on Adolescent SportsBehavior Econometric models are constructed to test therelationships between different social networks and ado-lescent sports behavior

Sporti βXi + PQ + e1 (10)

Among them Sport is adolescent sports behavior X isan adolescentsrsquo social network that includes an onlineofflinesocial network and an instrumentalemotionalmixed social

4 Complexity

network Q is the control variable matrix and e1 is theregression residual +e empirical results are shown inTable 2

+e results in Table 2 indicate that adolescentsrsquo socialnetworks have a significant positive effect on their sportsbehavior (β 0860 plt 0001) +e wider the homoge-neity of these adolescentsrsquo social relationship is the morefrequent their sports behavior becomes Online socialnetworks and offline social networks have positive effectson adolescent sports behavior respectively Howeverwhen adolescents have both an offline social network andan online social network the offline social network has apositive effect on adolescent sports (β 1381 plt 0001)while the online social network has a negative effect onadolescent sports (β 0517 plt 0001) An instrumentalnetwork emotional network and mixed networkhave positive effects on adolescent sports respectivelywith coefficients of 0735 0964 0880 and are significantat a 1 level However when adolescents have the threekinds of above social relations at the same time theemotional network and the mixed network still havepositive effects on adolescent sports but the influence ofthe instrumental network on adolescent sports then be-comes insignificant

322 Mediating Effect of Social Efficacy According to Wenand Ye [18] it is necessary to test the parameters of threeregression equations to verify the mediating effect

Sporti cSNi + PQ + e2 (11)

SEi αSNi + PQ + e3 (12)

Sporti cprimeSPi + bSEi + PQ + e4 (13)

Among these SN is social network SE is social efficacyQ is the control variable matrix and e2 sim e4 are the re-gression residuals Further c is the total effect of the in-dependent variable (social network) on the dependentvariable (adolescent sports behavior) a is the effect of socialnetwork on the intervening variable (social efficacy) b is theeffect of social efficacy on adolescent sports behavior aftercontrolling for the influence of social network cprime is the directeffect of social network on adolescent sports behavior aftercontrolling for the influence of social efficacy +e mediatingeffect is tested in five steps first test the coefficient c ofequation (11) if c is significant the intermediary effect issignificant otherwise there is a masking effect But whetherit is significant or not follow-up tests are carried out

1 2 3 4Number

Scree plot

000

050

100

150

200

250

300

350

Eige

nval

ues

(a)

1 2 3 4 5 6 7 8 9 10 11Number

Scree plot

000100200300400500600700800

Eige

nval

ues

(b)

Figure 1 Scree plot of adolescent sports (a) and social network (b)

Table 1 Mean standard deviation and correlation coefficient of variables

M SD 1 2 3 4 5 6 7 8 9 10 111 Age 1453 169 12 Gender 053 050 minus0018 13 Sports 288 093 minus0042 minus0076 14 Social network 208 087 minus0004lowast minus0105lowast 0858lowast 15 Online social network 209 089 0015lowast minus0084 0744lowastlowast 0967lowastlowast 16 Offline social network 236 064 minus0015 minus0113lowast 0903lowastlowast 0988lowastlowast 0916lowastlowast 17 Mixes social network 230 071 minus0018 minus0126lowastlowast 0876lowastlowast 0945lowast 0863lowastlowast 0966lowastlowast 18 Instrumental socialnetwork 234 066 0000 minus0107lowast 0734lowastlowast 0944lowastlowast 0923lowastlowast 0924lowastlowast 0823lowastlowast 1

9 Emotional socialnetwork 1453 169 minus0033 minus0066 0965lowastlowast 0844lowastlowast 0724lowastlowast 0891lowastlowast 0850lowastlowast 0711lowastlowast 1

10 Social efficacy 241 071 minus0058 minus0078 0980lowastlowast 0838lowastlowast 0725lowastlowast 0884lowastlowast 0880lowastlowast 0712lowastlowast 0916lowastlowast 111 Self-presentation 243 096 0006 minus0116lowastlowast 0808lowastlowast 0810lowastlowast 0710lowastlowast 0848lowastlowast 0895lowastlowast 0662lowastlowast 0816lowastlowast 0780lowastlowast 1Note N 568 gender is a dummy variable female 0 and male 1 lowastplt 005 lowastlowastplt 001

Complexity 5

Second test the coefficient a of equation (12) and the co-efficient b of equation (13) in turn if both are significantthen the indirect effect is significant go to the fourth step ifat least one is not significant go to the third step +ird useBootstrap method to directly check H0 ab 0 If it issignificant then the indirect effect is significant go to thefourth step otherwise the indirect effect is not significantstop the analysis Forth examining the coefficient cprime ofequation (13) if it is not significant the direct effect is not

significant indicating that there is only an intermediaryeffect If it is significant that is the direct effect is significantgo to the fifth step Fifth compare the signs of ab and cprime Ifthey are the same it is partly intermediary effect If the sign isdifferent it is a masking effect If a b and c are all significantthe mediation effect is significant otherwise there are othereffects For example if c is not significant there is a maskingeffect if cprime is significant there also exist direct effects Besidesthe above five steps a nonparametric percentile Bootstrap

Below 12 13-14 15-16 17-18

Statistics of gender in different ages

MaleFemale

010203040506070

()

(a)

Male Female

Statistics of ages in different genders

Below 1213-14

15-1617-18

05

101520253035404550

()

(b)

Online social networkOffline social networkMixed social network

Instrumental social networkEmotional social network

Below 12 13-14 15-16 17-18

Statistics of social networks in different ages

0

1

2

3

4

5

6

(c)

Once a weekOne to three times a weekFour to five times a weekMore than five times a week

Below 12 13-14 15-16 17-18

Statistics of sport behaviors in different ages

05

10152025303540

()

(d)

Figure 2 Statistic of adolescent social networks and sports behavior

Table 2 Estimated results of the influence of social network on adolescent sports behavior

Variables Adolescent sports behaviorConstant 0316 (1568) 0471lowast (1794) 0210 (1240) 0069 (0465) 0357 (1336) 0100 (0966) 0179 (0945) 0102 (1071)Age minus0023lowast (minus1667) minus0031lowast (-1767) minus0016 (minus1427) minus0007 (minus0737) minus0025 (minus1366) minus0006 (minus0859) minus0015 (minus1157) minus0007 (minus1113)Gender 0028 (0599) minus0028 (minus0458) 0053 (1356) 0074lowastlowast (2154) 0004 (0068) minus0024 (minus1004) 0070 (1596) 0004 (0203)Socialnetwork 0860lowastlowastlowast (37052) mdash mdash mdash mdash mdash mdash mdash

Online socialnetwork mdash 0744lowastlowastlowast (24673) mdash minus0517lowastlowastlowast (minus12194) mdash mdash mdash mdash

Offline socialnetwork mdash mdash 0905lowastlowastlowast (45545) 1381lowastlowastlowast (32455) mdash mdash mdash mdash

Instrumentalnetwork mdash mdash mdash mdash 0735lowastlowastlowast (23892) mdash mdash 0016 (0818)

Emotionalnetwork mdash mdash mdash mdash mdash 0964lowastlowastlowast (81163) mdash 0793lowastlowastlowast (38244)

Mixednetwork mdash mdash mdash mdash mdash mdash 0880lowastlowastlowast (40217) 0189lowastlowastlowast (7352)

Adj-R2 0737 0554 0816 0858 0538 0931 0767 0942

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

6 Complexity

method was used to test the mediating effect If the 95confidence interval does not contain a zero then the me-diating effect becomes significant +e empirical regressionresults and significance test results for the mediating effectare shown in Table 3

+e results in Table 3 show that R2 in the three equationswas bigger than 07 and the goodness of fit for the model wasgood In equation (11) social network has a significantpositive effect on adolescent sports behaviorc 0860 plt 0001 In equation (12) social network has asignificant effect on social efficacy c 0839 plt 0001 Inequation (13) social network and social efficacy have asignificant effect on adolescent sports behaviorcprime 0122 b 0879 plt 0001 According to the mediatingeffect test procedure of Wen and Ye [18] the mediatingeffect of social efficacy is significant because a b c and cprime areall significant Further the Bootstrap method was used totest the mediating effect of social efficacy 5000 Bootstrapsamples were randomly selected from the original sample forindirect effect estimation Table 3 shows the estimated valueof the indirect effect of social efficacy to be 07374 and 95confidence intervals were [07007 07704] excluding zero+us the mediating effect of social efficacy becamesignificant

323 Moderating Effect of Self-Presentation As for themoderating effect of self-presentation the followingeconometric model was constructed

Sporti αSNi + βSPi + cSNi lowast SPi + PQ + e5 (14)

where SP is the self-presentation +e stepwise regressionmethod was used to estimate the moderating effect of self-presentation +e results are shown in Table 4

+e results of the study show that although both socialnetworking and self-presentation have significant positiveeffects on adolescent sports behavior the interaction termsof social network and self-presentation do not have anysignificant influence on adolescent sports behavior therebyindicating that the moderating effect of self-presentation isnot significant

In addition this study found that age has a negativepredictive effect on adolescent sports behavior that is theolder the adolescents are the less likely they are to exerciseGender had no significant effect on the prediction of ado-lescent sports behavior

33 Discussion

331 Influence of Social Networks on Adolescent SportsBehavior +e results of this study indicate that socialnetworks have significant influences on adolescent sportsbehavior demonstrating that adolescents with larger anddenser social networks have more frequent sports behaviorPeople are mutually linked to each other and so are theirsports behaviors +us the existence of social networksmeans that personsrsquo movements are interdependent [19]Sports behavior can be spread from person to person and

there is also a phenomenon called ldquopeer effectsrdquo [20] whichis just as the proverb says ldquoIf you live with a lame person youwill learn to limprdquo However it is interesting that with thestrengthening of the social network establishment the effectsof online social networks and offline social networks onadolescent sports behavior are not the same When ado-lescents have both an offline social network and an onlinesocial network the offline social network has a positive effecton adolescent sports while the online social network has anegative effect on adolescent sports +is outcome may bebecause at the present stage Chinese adolescents makefriends online mainly through games chats and other ways[21] Moreover adolescents in junior and senior highschools have heavy daily learning tasks and usually surf theInternet for leisure and entertainment +erefore mostadolescentsrsquo online social network relationships are enter-tainment groups Besides during the formation of onlinenetwork relationships most adolescents will use computersor mobile phones for a long time a factor that further re-duces the possibility of adolescent sports [22] Under theinfluence of the characteristics of network groups and thetime used for online communication offline social networkshave a negative influence on adolescent sports behavior

For social networks with different characteristics the in-strumental network emotional network and mixed networkhave positive effects on adolescent sports respectively In otherwords without the influence of others social relations such asthose with counselors teachers family members and friendshave positive effects on adolescent sports behaviors Howeverwhen adolescents have the above three kinds of social relationsat the same time the influence of the instrumental networkthat is counselors or teachers on adolescent sports behavior isnot significant +at is to say the influence of counselors orteachers on adolescent sports behavior is easily affected byother people +is outcome may be because the instrumentalnetwork connectivity is not strong [23] Even for teachers andstudents who live and study very closely there is no similarinfluence like those of friends indicating that although physicalactivity has the interpersonal infectivity of social network itsinfectivity changes based on relative activities and the gendersof friends [24] For example in a given social network lessactive runners tend to affect more active runners while theopposite is not true Both men and women affect men whileonly women affect other women [25]+e reason why it is easyfor physical activity behaviors to spread through adolescentfriendship or a family network is that first of all adolescents aremore willing to experience peer relations during adolescenceand their peer groups are based on common behaviors (sportscomputer games video games diet etc) +ese directly orindirectly affect their physical activities and health [26] Sec-ondly family relationship is the network relationship with thestrongest connection among adolescents Under the influenceof that strong connection an emotional network will have anobvious influence on adolescent sports behavior

Overall different types of social networks have positiveeffects on adolescent sports behavior separately but differentsocial networks are also interacted so their influences onadolescent sports behavior will change by the interactions

Complexity 7

+at is to say the effect of social networks on sports behaviornot only is influenced by the strength and density of socialnetworks but also influenced the interaction of differentsocial networks

332 Mediating Effect of Social Efficacy and the ModeratingEffect of Self-Presentation Social efficacy plays a mediatingrole in the relationship between social networking andadolescent sports behavior and these results can beexplained using the following perspectives

First according to the theoretical model of ldquotwo pathsand three layersrdquo of trust behavior social relationship is asituational factor and its influence on adolescent sportsbehavior is affected by internal cognition [27] In a positiveand extroverted social relationship adolescent sports be-haviors tend to receive more social support which then canenable adolescents to form a more positive internal cogni-tion of themselves and others thereby producing a strongermotivation to undertake sports [28]

Secondly from the perspective of self-efficacy theorysituational condition is one of the important factors that canaffect individual self-efficacy [29] If one is in a familiar andpleasant environment individualrsquos self-efficacy will increaseDoing sports is an important way to relieve adolescentpsychological stress and improve physical health +usadolescents receive positive demonstrations and responsesfrom their social network and their motivation to exercisewill strengthen in order to achieve the recognition of theirsocial relationships

+e moderating effect of self-presentation in the rela-tionship between a social network and adolescent sportsbehavior is not significant in that the positive effect of social

network on adolescent sports behavior is not affected byadolescentsrsquo self-presentation+is outcomemay be becauseon the one hand the level of self-presentation in Chineseadolescentsrsquo sports is low According to the Research Reporton Internet Usersrsquo Information Security released by the ChinaInternet Network Information Center (CNNIC 2018) morethan 73 of adolescentsrsquo self-presentation focuses on en-tertainment and study not so much on sports [30] On theother hand although self-presentation can reduce theloneliness of individuals and show them the enthusiasm ofsports behaviors it also presents certain negative informa-tion about sports such as fatigue and spending time therebyweakening the positive effect of self-presentation [31]

To sum up then this study concludes that social net-working not only directly predicts adolescent sports be-havior but also indirectly affects their actual sports behaviorthrough social efficacy +e implication of this mediatormodel is that adolescent sports behavior can also be gen-erated through interpersonal interaction and that behaviorcan be affected by individual cognition thereby showing therelevance of self-cognition gained by social interaction [32]

4 Conclusions

+e main conclusions of this study are the following

(1) +e social networks of adolescents have obviouspositive effects on their sports behavior Single onlinesocial networks and offline social networks instru-mental networks emotional networks and mixednetworks have obvious positive effects on adolescentsports behavior respectively However under theinfluence of multiple types of social networks offline

Table 4 Estimated results of moderating effect of self-presentation

Variables Adolescent sports behaviorConstant 0383lowast (1654) 0330lowast (1768) 0317lowast (1691)Age minus0028lowast (minus1764) minus0024lowast (minus1936) minus0024lowast (minus1916)Gender 0036 (0673) 0049 (1134) 0046 (1069)Self- presentation 0811lowastlowastlowast (30426) 0332lowastlowastlowast (9093) 0329lowastlowastlowast (8993)Social network mdash 0592lowastlowastlowast (16231) 0589lowastlowastlowast (16109)Social networklowastself-presentation mdash mdash 0014 (0891)Adj-R2 0654 0774 0774Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

Table 3 Estimated results and significance test results of mediating effect of social efficacy

Variables Equation (11)Sports

Equation (12)Social efficacy

Equation (13)Sports

Constant 0316 (1568) 0460lowastlowast (2148) minus0088 (minus1191)Age minus0023lowast (minus1667) minus0032lowastlowast (minus2228) 0006 (1123)Gender 0028 (0599) 0018 (0364) 0012 (0716)Social network 0860lowastlowastlowast (37052) 0839lowastlowastlowast (34099) 0122lowastlowastlowast (7916)Social efficacy mdash mdash 0879lowastlowastlowast (5691)Adj-R2 0737 0706 0956Indirect effect(s) of SN on sport

Intervening variable social contact Effect Boot standard error Boot LLCI Boot ULCI07374 00178 07007 07704

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

8 Complexity

social networking has a negative predictive effect onadolescent sports behavior while the effect of mixednetworks is not so obvious

(2) Social efficacy plays an intermediary role in the re-lationship between social networking and adolescentsports behavior +e moderating effect of self-pre-sentation is not significant

Although this paper found the importance of socialnetworks in adolescents sports behavior and discovered therole of social efficacy there are some limitations to this studywhich need to be addressed in future studies For example allvariables in this study were investigated using a questionnaireit canmeasure variables in static level butmay ignore diversityand complexity of variables furthermore this study foundthat the moderating effect of self-presentation is not signif-icant+is finding may have a theoretical rationale but it mayalso be due to the low degree of self-presentation of thesample In the future that sample needs to be expanded forfurther research and more precise evaluation

Data Availability

+e survey data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

+is study was supported by the MOE (Ministry of Edu-cation in China) Project of Humanities and Social Sciences(Project no 18YJC890012)

References

[1] L P Spear ldquo+e adolescent brain and age-related behavioralmanifestationsrdquo Neuroscience amp Biobehavioral Reviewsvol 24 no 4 pp 417ndash463 2000

[2] L Steinberg and K C Monahan ldquoAge differences in resis-tance to peer influencerdquo Developmental Psychology vol 43no 6 pp 1531ndash1543 2009

[3] S Sarah-Jeanne N R James C B Julie et al ldquoEffect of peersand friends on youth physical activity and motivation to bephysically activerdquo Journal of Pediatric Psychology vol 34no 2 pp 217ndash225 2009

[4] M Jennifer D L H Kayla L M Barnett et al ldquoFriendshipnetwork characteristics are associated with physical activityand sedentary behavior in early adolescencerdquo PLoS ONEvol 10 no 12 pp 1ndash15 2015

[5] M J Edwards R Jago J S Simon et al ldquo+e influence offriends and siblings on the physical activity and screenviewing behaviours of children aged 5-6 years a qualitativeanalysis of parent interviewsrdquo BMJ Open vol 5 no 5 ArticleID e006593 2015

[6] K D L Haye G Robins PMohr et al ldquoHow physical activityshapes and is shaped by adolescent friendshipsrdquo SocialScience amp Medicine vol 73 no 5 pp 719ndash728 2011

[7] Z Li F Xiong X Wang et al ldquoTopological influence-awarerecommendation on social networksrdquo Complexity vol 2019Article ID 6325654 12 pages 2019

[8] B Yu and X Hu ldquoToward training and assessing reproducibledata analysis in data science educationrdquo Data Intelligencevol 1 no 4 pp 333ndash344 2019

[9] A Bandura ldquoSelf-efficacy mechanism in human agencyrdquoAmerican Psychologist vol 37 no 2 pp 122ndash147 1982

[10] A Bandura Self-efficacy He Exercise of Control WorthPublishers New Nork NY USA 1997

[11] C Fan and A S Mak ldquoMeasuring social self-efficacy in aculturally diverse student populationrdquo Social Behavior andPersonality An International Journal vol 26 no 2 pp 131ndash144 1998

[12] M Wei D W Russell and R A Zakalik ldquoAdult attachmentsocial self-efficacy self-disclosure loneliness and subsequentdepression for freshman college students a longitudinal studyrdquoJournal of Counseling Psychology vol 52 no 4 pp 602ndash614 2005

[13] F Xiong W Shen H Chen et al ldquoExploiting implicit in-fluence from information propagation for social recom-mendationrdquo IEEE Transactions on Cybernetics 2019

[14] H Y Mao H C Hsu and S D Lee ldquoGender differences inrelated influential factors of regular exercise behavior amongpeople in Taiwan in 2007 a cross-sectional studyrdquo PLoS onevol 15 no 1 pp 1ndash22 2020

[15] S Park J Y Kang and L A Chadiha ldquoSocial network typeshealth and health-Care use among South Korean olderadultsrdquo Research on Aging vol 40 no 2 pp 131ndash154 2018

[16] E J Jeong and D H Kim ldquoSocial activities self-efficacy gameattitudes and game addictionrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 4 pp 213ndash221 2011

[17] J Kim and J-E R Lee ldquo+e Facebook paths to happinesseffects of the number of Facebook friends and self-presen-tation on subjective well-beingrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 6 pp 359ndash364 2011

[18] Z Wen and B Ye ldquoAnalyses of mediating effects the de-velopment of methods and modelsrdquo Advances in Psycho-logical Science vol 22 no 5 pp 731ndash745 2014

[19] L G Smith L Banting R Eime et al ldquo+e association be-tween social support and physical activity in older adultsasystematic reviewrdquo International Journal of Behavioral Nu-trition amp Physical Activity vol 14 no 1 p 56 2017

[20] K P Smith and N A Christakis ldquoSocial networks andhealthrdquo Annual Review of Sociology vol 34 no 1pp 405ndash429 2008

[21] S Aral and C Nicolaides ldquoExercise contagion in a globalsocial networkrdquo Nature Communications vol 8 Article ID14753 2017

[22] S Ji S Pan E Cambria et al ldquoA survey on knowledge graphsrepresentation acquisition and applicationsrdquo 2020 httparxivorgabs200200388

[23] T W Valente K Fujimoto C-P Chou and D Spruijt-MetzldquoAdolescent affiliations and adiposity a social networkanalysis of friendships and obesityrdquo Journal of AdolescentHealth vol 45 no 2 pp 202ndash204 2009

[24] P Manasatchakun P Chotiga J Hochwalder A RoxbergM Sandborgh and M Asp ldquoFactors associated with healthyaging among older persons in northeastern+ailandrdquo Journalof Cross-Cultural Gerontology vol 31 no 4 pp 369ndash3842016

[25] L Adam and P S Damon ldquo+e intersection of sport man-agement and sociology of sport research a social networkperspectiverdquo Sport Management Review vol 15 no 2pp 244ndash256 2012

Complexity 9

network Q is the control variable matrix and e1 is theregression residual +e empirical results are shown inTable 2

+e results in Table 2 indicate that adolescentsrsquo socialnetworks have a significant positive effect on their sportsbehavior (β 0860 plt 0001) +e wider the homoge-neity of these adolescentsrsquo social relationship is the morefrequent their sports behavior becomes Online socialnetworks and offline social networks have positive effectson adolescent sports behavior respectively Howeverwhen adolescents have both an offline social network andan online social network the offline social network has apositive effect on adolescent sports (β 1381 plt 0001)while the online social network has a negative effect onadolescent sports (β 0517 plt 0001) An instrumentalnetwork emotional network and mixed networkhave positive effects on adolescent sports respectivelywith coefficients of 0735 0964 0880 and are significantat a 1 level However when adolescents have the threekinds of above social relations at the same time theemotional network and the mixed network still havepositive effects on adolescent sports but the influence ofthe instrumental network on adolescent sports then be-comes insignificant

322 Mediating Effect of Social Efficacy According to Wenand Ye [18] it is necessary to test the parameters of threeregression equations to verify the mediating effect

Sporti cSNi + PQ + e2 (11)

SEi αSNi + PQ + e3 (12)

Sporti cprimeSPi + bSEi + PQ + e4 (13)

Among these SN is social network SE is social efficacyQ is the control variable matrix and e2 sim e4 are the re-gression residuals Further c is the total effect of the in-dependent variable (social network) on the dependentvariable (adolescent sports behavior) a is the effect of socialnetwork on the intervening variable (social efficacy) b is theeffect of social efficacy on adolescent sports behavior aftercontrolling for the influence of social network cprime is the directeffect of social network on adolescent sports behavior aftercontrolling for the influence of social efficacy +e mediatingeffect is tested in five steps first test the coefficient c ofequation (11) if c is significant the intermediary effect issignificant otherwise there is a masking effect But whetherit is significant or not follow-up tests are carried out

1 2 3 4Number

Scree plot

000

050

100

150

200

250

300

350

Eige

nval

ues

(a)

1 2 3 4 5 6 7 8 9 10 11Number

Scree plot

000100200300400500600700800

Eige

nval

ues

(b)

Figure 1 Scree plot of adolescent sports (a) and social network (b)

Table 1 Mean standard deviation and correlation coefficient of variables

M SD 1 2 3 4 5 6 7 8 9 10 111 Age 1453 169 12 Gender 053 050 minus0018 13 Sports 288 093 minus0042 minus0076 14 Social network 208 087 minus0004lowast minus0105lowast 0858lowast 15 Online social network 209 089 0015lowast minus0084 0744lowastlowast 0967lowastlowast 16 Offline social network 236 064 minus0015 minus0113lowast 0903lowastlowast 0988lowastlowast 0916lowastlowast 17 Mixes social network 230 071 minus0018 minus0126lowastlowast 0876lowastlowast 0945lowast 0863lowastlowast 0966lowastlowast 18 Instrumental socialnetwork 234 066 0000 minus0107lowast 0734lowastlowast 0944lowastlowast 0923lowastlowast 0924lowastlowast 0823lowastlowast 1

9 Emotional socialnetwork 1453 169 minus0033 minus0066 0965lowastlowast 0844lowastlowast 0724lowastlowast 0891lowastlowast 0850lowastlowast 0711lowastlowast 1

10 Social efficacy 241 071 minus0058 minus0078 0980lowastlowast 0838lowastlowast 0725lowastlowast 0884lowastlowast 0880lowastlowast 0712lowastlowast 0916lowastlowast 111 Self-presentation 243 096 0006 minus0116lowastlowast 0808lowastlowast 0810lowastlowast 0710lowastlowast 0848lowastlowast 0895lowastlowast 0662lowastlowast 0816lowastlowast 0780lowastlowast 1Note N 568 gender is a dummy variable female 0 and male 1 lowastplt 005 lowastlowastplt 001

Complexity 5

Second test the coefficient a of equation (12) and the co-efficient b of equation (13) in turn if both are significantthen the indirect effect is significant go to the fourth step ifat least one is not significant go to the third step +ird useBootstrap method to directly check H0 ab 0 If it issignificant then the indirect effect is significant go to thefourth step otherwise the indirect effect is not significantstop the analysis Forth examining the coefficient cprime ofequation (13) if it is not significant the direct effect is not

significant indicating that there is only an intermediaryeffect If it is significant that is the direct effect is significantgo to the fifth step Fifth compare the signs of ab and cprime Ifthey are the same it is partly intermediary effect If the sign isdifferent it is a masking effect If a b and c are all significantthe mediation effect is significant otherwise there are othereffects For example if c is not significant there is a maskingeffect if cprime is significant there also exist direct effects Besidesthe above five steps a nonparametric percentile Bootstrap

Below 12 13-14 15-16 17-18

Statistics of gender in different ages

MaleFemale

010203040506070

()

(a)

Male Female

Statistics of ages in different genders

Below 1213-14

15-1617-18

05

101520253035404550

()

(b)

Online social networkOffline social networkMixed social network

Instrumental social networkEmotional social network

Below 12 13-14 15-16 17-18

Statistics of social networks in different ages

0

1

2

3

4

5

6

(c)

Once a weekOne to three times a weekFour to five times a weekMore than five times a week

Below 12 13-14 15-16 17-18

Statistics of sport behaviors in different ages

05

10152025303540

()

(d)

Figure 2 Statistic of adolescent social networks and sports behavior

Table 2 Estimated results of the influence of social network on adolescent sports behavior

Variables Adolescent sports behaviorConstant 0316 (1568) 0471lowast (1794) 0210 (1240) 0069 (0465) 0357 (1336) 0100 (0966) 0179 (0945) 0102 (1071)Age minus0023lowast (minus1667) minus0031lowast (-1767) minus0016 (minus1427) minus0007 (minus0737) minus0025 (minus1366) minus0006 (minus0859) minus0015 (minus1157) minus0007 (minus1113)Gender 0028 (0599) minus0028 (minus0458) 0053 (1356) 0074lowastlowast (2154) 0004 (0068) minus0024 (minus1004) 0070 (1596) 0004 (0203)Socialnetwork 0860lowastlowastlowast (37052) mdash mdash mdash mdash mdash mdash mdash

Online socialnetwork mdash 0744lowastlowastlowast (24673) mdash minus0517lowastlowastlowast (minus12194) mdash mdash mdash mdash

Offline socialnetwork mdash mdash 0905lowastlowastlowast (45545) 1381lowastlowastlowast (32455) mdash mdash mdash mdash

Instrumentalnetwork mdash mdash mdash mdash 0735lowastlowastlowast (23892) mdash mdash 0016 (0818)

Emotionalnetwork mdash mdash mdash mdash mdash 0964lowastlowastlowast (81163) mdash 0793lowastlowastlowast (38244)

Mixednetwork mdash mdash mdash mdash mdash mdash 0880lowastlowastlowast (40217) 0189lowastlowastlowast (7352)

Adj-R2 0737 0554 0816 0858 0538 0931 0767 0942

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

6 Complexity

method was used to test the mediating effect If the 95confidence interval does not contain a zero then the me-diating effect becomes significant +e empirical regressionresults and significance test results for the mediating effectare shown in Table 3

+e results in Table 3 show that R2 in the three equationswas bigger than 07 and the goodness of fit for the model wasgood In equation (11) social network has a significantpositive effect on adolescent sports behaviorc 0860 plt 0001 In equation (12) social network has asignificant effect on social efficacy c 0839 plt 0001 Inequation (13) social network and social efficacy have asignificant effect on adolescent sports behaviorcprime 0122 b 0879 plt 0001 According to the mediatingeffect test procedure of Wen and Ye [18] the mediatingeffect of social efficacy is significant because a b c and cprime areall significant Further the Bootstrap method was used totest the mediating effect of social efficacy 5000 Bootstrapsamples were randomly selected from the original sample forindirect effect estimation Table 3 shows the estimated valueof the indirect effect of social efficacy to be 07374 and 95confidence intervals were [07007 07704] excluding zero+us the mediating effect of social efficacy becamesignificant

323 Moderating Effect of Self-Presentation As for themoderating effect of self-presentation the followingeconometric model was constructed

Sporti αSNi + βSPi + cSNi lowast SPi + PQ + e5 (14)

where SP is the self-presentation +e stepwise regressionmethod was used to estimate the moderating effect of self-presentation +e results are shown in Table 4

+e results of the study show that although both socialnetworking and self-presentation have significant positiveeffects on adolescent sports behavior the interaction termsof social network and self-presentation do not have anysignificant influence on adolescent sports behavior therebyindicating that the moderating effect of self-presentation isnot significant

In addition this study found that age has a negativepredictive effect on adolescent sports behavior that is theolder the adolescents are the less likely they are to exerciseGender had no significant effect on the prediction of ado-lescent sports behavior

33 Discussion

331 Influence of Social Networks on Adolescent SportsBehavior +e results of this study indicate that socialnetworks have significant influences on adolescent sportsbehavior demonstrating that adolescents with larger anddenser social networks have more frequent sports behaviorPeople are mutually linked to each other and so are theirsports behaviors +us the existence of social networksmeans that personsrsquo movements are interdependent [19]Sports behavior can be spread from person to person and

there is also a phenomenon called ldquopeer effectsrdquo [20] whichis just as the proverb says ldquoIf you live with a lame person youwill learn to limprdquo However it is interesting that with thestrengthening of the social network establishment the effectsof online social networks and offline social networks onadolescent sports behavior are not the same When ado-lescents have both an offline social network and an onlinesocial network the offline social network has a positive effecton adolescent sports while the online social network has anegative effect on adolescent sports +is outcome may bebecause at the present stage Chinese adolescents makefriends online mainly through games chats and other ways[21] Moreover adolescents in junior and senior highschools have heavy daily learning tasks and usually surf theInternet for leisure and entertainment +erefore mostadolescentsrsquo online social network relationships are enter-tainment groups Besides during the formation of onlinenetwork relationships most adolescents will use computersor mobile phones for a long time a factor that further re-duces the possibility of adolescent sports [22] Under theinfluence of the characteristics of network groups and thetime used for online communication offline social networkshave a negative influence on adolescent sports behavior

For social networks with different characteristics the in-strumental network emotional network and mixed networkhave positive effects on adolescent sports respectively In otherwords without the influence of others social relations such asthose with counselors teachers family members and friendshave positive effects on adolescent sports behaviors Howeverwhen adolescents have the above three kinds of social relationsat the same time the influence of the instrumental networkthat is counselors or teachers on adolescent sports behavior isnot significant +at is to say the influence of counselors orteachers on adolescent sports behavior is easily affected byother people +is outcome may be because the instrumentalnetwork connectivity is not strong [23] Even for teachers andstudents who live and study very closely there is no similarinfluence like those of friends indicating that although physicalactivity has the interpersonal infectivity of social network itsinfectivity changes based on relative activities and the gendersof friends [24] For example in a given social network lessactive runners tend to affect more active runners while theopposite is not true Both men and women affect men whileonly women affect other women [25]+e reason why it is easyfor physical activity behaviors to spread through adolescentfriendship or a family network is that first of all adolescents aremore willing to experience peer relations during adolescenceand their peer groups are based on common behaviors (sportscomputer games video games diet etc) +ese directly orindirectly affect their physical activities and health [26] Sec-ondly family relationship is the network relationship with thestrongest connection among adolescents Under the influenceof that strong connection an emotional network will have anobvious influence on adolescent sports behavior

Overall different types of social networks have positiveeffects on adolescent sports behavior separately but differentsocial networks are also interacted so their influences onadolescent sports behavior will change by the interactions

Complexity 7

+at is to say the effect of social networks on sports behaviornot only is influenced by the strength and density of socialnetworks but also influenced the interaction of differentsocial networks

332 Mediating Effect of Social Efficacy and the ModeratingEffect of Self-Presentation Social efficacy plays a mediatingrole in the relationship between social networking andadolescent sports behavior and these results can beexplained using the following perspectives

First according to the theoretical model of ldquotwo pathsand three layersrdquo of trust behavior social relationship is asituational factor and its influence on adolescent sportsbehavior is affected by internal cognition [27] In a positiveand extroverted social relationship adolescent sports be-haviors tend to receive more social support which then canenable adolescents to form a more positive internal cogni-tion of themselves and others thereby producing a strongermotivation to undertake sports [28]

Secondly from the perspective of self-efficacy theorysituational condition is one of the important factors that canaffect individual self-efficacy [29] If one is in a familiar andpleasant environment individualrsquos self-efficacy will increaseDoing sports is an important way to relieve adolescentpsychological stress and improve physical health +usadolescents receive positive demonstrations and responsesfrom their social network and their motivation to exercisewill strengthen in order to achieve the recognition of theirsocial relationships

+e moderating effect of self-presentation in the rela-tionship between a social network and adolescent sportsbehavior is not significant in that the positive effect of social

network on adolescent sports behavior is not affected byadolescentsrsquo self-presentation+is outcomemay be becauseon the one hand the level of self-presentation in Chineseadolescentsrsquo sports is low According to the Research Reporton Internet Usersrsquo Information Security released by the ChinaInternet Network Information Center (CNNIC 2018) morethan 73 of adolescentsrsquo self-presentation focuses on en-tertainment and study not so much on sports [30] On theother hand although self-presentation can reduce theloneliness of individuals and show them the enthusiasm ofsports behaviors it also presents certain negative informa-tion about sports such as fatigue and spending time therebyweakening the positive effect of self-presentation [31]

To sum up then this study concludes that social net-working not only directly predicts adolescent sports be-havior but also indirectly affects their actual sports behaviorthrough social efficacy +e implication of this mediatormodel is that adolescent sports behavior can also be gen-erated through interpersonal interaction and that behaviorcan be affected by individual cognition thereby showing therelevance of self-cognition gained by social interaction [32]

4 Conclusions

+e main conclusions of this study are the following

(1) +e social networks of adolescents have obviouspositive effects on their sports behavior Single onlinesocial networks and offline social networks instru-mental networks emotional networks and mixednetworks have obvious positive effects on adolescentsports behavior respectively However under theinfluence of multiple types of social networks offline

Table 4 Estimated results of moderating effect of self-presentation

Variables Adolescent sports behaviorConstant 0383lowast (1654) 0330lowast (1768) 0317lowast (1691)Age minus0028lowast (minus1764) minus0024lowast (minus1936) minus0024lowast (minus1916)Gender 0036 (0673) 0049 (1134) 0046 (1069)Self- presentation 0811lowastlowastlowast (30426) 0332lowastlowastlowast (9093) 0329lowastlowastlowast (8993)Social network mdash 0592lowastlowastlowast (16231) 0589lowastlowastlowast (16109)Social networklowastself-presentation mdash mdash 0014 (0891)Adj-R2 0654 0774 0774Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

Table 3 Estimated results and significance test results of mediating effect of social efficacy

Variables Equation (11)Sports

Equation (12)Social efficacy

Equation (13)Sports

Constant 0316 (1568) 0460lowastlowast (2148) minus0088 (minus1191)Age minus0023lowast (minus1667) minus0032lowastlowast (minus2228) 0006 (1123)Gender 0028 (0599) 0018 (0364) 0012 (0716)Social network 0860lowastlowastlowast (37052) 0839lowastlowastlowast (34099) 0122lowastlowastlowast (7916)Social efficacy mdash mdash 0879lowastlowastlowast (5691)Adj-R2 0737 0706 0956Indirect effect(s) of SN on sport

Intervening variable social contact Effect Boot standard error Boot LLCI Boot ULCI07374 00178 07007 07704

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

8 Complexity

social networking has a negative predictive effect onadolescent sports behavior while the effect of mixednetworks is not so obvious

(2) Social efficacy plays an intermediary role in the re-lationship between social networking and adolescentsports behavior +e moderating effect of self-pre-sentation is not significant

Although this paper found the importance of socialnetworks in adolescents sports behavior and discovered therole of social efficacy there are some limitations to this studywhich need to be addressed in future studies For example allvariables in this study were investigated using a questionnaireit canmeasure variables in static level butmay ignore diversityand complexity of variables furthermore this study foundthat the moderating effect of self-presentation is not signif-icant+is finding may have a theoretical rationale but it mayalso be due to the low degree of self-presentation of thesample In the future that sample needs to be expanded forfurther research and more precise evaluation

Data Availability

+e survey data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

+is study was supported by the MOE (Ministry of Edu-cation in China) Project of Humanities and Social Sciences(Project no 18YJC890012)

References

[1] L P Spear ldquo+e adolescent brain and age-related behavioralmanifestationsrdquo Neuroscience amp Biobehavioral Reviewsvol 24 no 4 pp 417ndash463 2000

[2] L Steinberg and K C Monahan ldquoAge differences in resis-tance to peer influencerdquo Developmental Psychology vol 43no 6 pp 1531ndash1543 2009

[3] S Sarah-Jeanne N R James C B Julie et al ldquoEffect of peersand friends on youth physical activity and motivation to bephysically activerdquo Journal of Pediatric Psychology vol 34no 2 pp 217ndash225 2009

[4] M Jennifer D L H Kayla L M Barnett et al ldquoFriendshipnetwork characteristics are associated with physical activityand sedentary behavior in early adolescencerdquo PLoS ONEvol 10 no 12 pp 1ndash15 2015

[5] M J Edwards R Jago J S Simon et al ldquo+e influence offriends and siblings on the physical activity and screenviewing behaviours of children aged 5-6 years a qualitativeanalysis of parent interviewsrdquo BMJ Open vol 5 no 5 ArticleID e006593 2015

[6] K D L Haye G Robins PMohr et al ldquoHow physical activityshapes and is shaped by adolescent friendshipsrdquo SocialScience amp Medicine vol 73 no 5 pp 719ndash728 2011

[7] Z Li F Xiong X Wang et al ldquoTopological influence-awarerecommendation on social networksrdquo Complexity vol 2019Article ID 6325654 12 pages 2019

[8] B Yu and X Hu ldquoToward training and assessing reproducibledata analysis in data science educationrdquo Data Intelligencevol 1 no 4 pp 333ndash344 2019

[9] A Bandura ldquoSelf-efficacy mechanism in human agencyrdquoAmerican Psychologist vol 37 no 2 pp 122ndash147 1982

[10] A Bandura Self-efficacy He Exercise of Control WorthPublishers New Nork NY USA 1997

[11] C Fan and A S Mak ldquoMeasuring social self-efficacy in aculturally diverse student populationrdquo Social Behavior andPersonality An International Journal vol 26 no 2 pp 131ndash144 1998

[12] M Wei D W Russell and R A Zakalik ldquoAdult attachmentsocial self-efficacy self-disclosure loneliness and subsequentdepression for freshman college students a longitudinal studyrdquoJournal of Counseling Psychology vol 52 no 4 pp 602ndash614 2005

[13] F Xiong W Shen H Chen et al ldquoExploiting implicit in-fluence from information propagation for social recom-mendationrdquo IEEE Transactions on Cybernetics 2019

[14] H Y Mao H C Hsu and S D Lee ldquoGender differences inrelated influential factors of regular exercise behavior amongpeople in Taiwan in 2007 a cross-sectional studyrdquo PLoS onevol 15 no 1 pp 1ndash22 2020

[15] S Park J Y Kang and L A Chadiha ldquoSocial network typeshealth and health-Care use among South Korean olderadultsrdquo Research on Aging vol 40 no 2 pp 131ndash154 2018

[16] E J Jeong and D H Kim ldquoSocial activities self-efficacy gameattitudes and game addictionrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 4 pp 213ndash221 2011

[17] J Kim and J-E R Lee ldquo+e Facebook paths to happinesseffects of the number of Facebook friends and self-presen-tation on subjective well-beingrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 6 pp 359ndash364 2011

[18] Z Wen and B Ye ldquoAnalyses of mediating effects the de-velopment of methods and modelsrdquo Advances in Psycho-logical Science vol 22 no 5 pp 731ndash745 2014

[19] L G Smith L Banting R Eime et al ldquo+e association be-tween social support and physical activity in older adultsasystematic reviewrdquo International Journal of Behavioral Nu-trition amp Physical Activity vol 14 no 1 p 56 2017

[20] K P Smith and N A Christakis ldquoSocial networks andhealthrdquo Annual Review of Sociology vol 34 no 1pp 405ndash429 2008

[21] S Aral and C Nicolaides ldquoExercise contagion in a globalsocial networkrdquo Nature Communications vol 8 Article ID14753 2017

[22] S Ji S Pan E Cambria et al ldquoA survey on knowledge graphsrepresentation acquisition and applicationsrdquo 2020 httparxivorgabs200200388

[23] T W Valente K Fujimoto C-P Chou and D Spruijt-MetzldquoAdolescent affiliations and adiposity a social networkanalysis of friendships and obesityrdquo Journal of AdolescentHealth vol 45 no 2 pp 202ndash204 2009

[24] P Manasatchakun P Chotiga J Hochwalder A RoxbergM Sandborgh and M Asp ldquoFactors associated with healthyaging among older persons in northeastern+ailandrdquo Journalof Cross-Cultural Gerontology vol 31 no 4 pp 369ndash3842016

[25] L Adam and P S Damon ldquo+e intersection of sport man-agement and sociology of sport research a social networkperspectiverdquo Sport Management Review vol 15 no 2pp 244ndash256 2012

Complexity 9

Second test the coefficient a of equation (12) and the co-efficient b of equation (13) in turn if both are significantthen the indirect effect is significant go to the fourth step ifat least one is not significant go to the third step +ird useBootstrap method to directly check H0 ab 0 If it issignificant then the indirect effect is significant go to thefourth step otherwise the indirect effect is not significantstop the analysis Forth examining the coefficient cprime ofequation (13) if it is not significant the direct effect is not

significant indicating that there is only an intermediaryeffect If it is significant that is the direct effect is significantgo to the fifth step Fifth compare the signs of ab and cprime Ifthey are the same it is partly intermediary effect If the sign isdifferent it is a masking effect If a b and c are all significantthe mediation effect is significant otherwise there are othereffects For example if c is not significant there is a maskingeffect if cprime is significant there also exist direct effects Besidesthe above five steps a nonparametric percentile Bootstrap

Below 12 13-14 15-16 17-18

Statistics of gender in different ages

MaleFemale

010203040506070

()

(a)

Male Female

Statistics of ages in different genders

Below 1213-14

15-1617-18

05

101520253035404550

()

(b)

Online social networkOffline social networkMixed social network

Instrumental social networkEmotional social network

Below 12 13-14 15-16 17-18

Statistics of social networks in different ages

0

1

2

3

4

5

6

(c)

Once a weekOne to three times a weekFour to five times a weekMore than five times a week

Below 12 13-14 15-16 17-18

Statistics of sport behaviors in different ages

05

10152025303540

()

(d)

Figure 2 Statistic of adolescent social networks and sports behavior

Table 2 Estimated results of the influence of social network on adolescent sports behavior

Variables Adolescent sports behaviorConstant 0316 (1568) 0471lowast (1794) 0210 (1240) 0069 (0465) 0357 (1336) 0100 (0966) 0179 (0945) 0102 (1071)Age minus0023lowast (minus1667) minus0031lowast (-1767) minus0016 (minus1427) minus0007 (minus0737) minus0025 (minus1366) minus0006 (minus0859) minus0015 (minus1157) minus0007 (minus1113)Gender 0028 (0599) minus0028 (minus0458) 0053 (1356) 0074lowastlowast (2154) 0004 (0068) minus0024 (minus1004) 0070 (1596) 0004 (0203)Socialnetwork 0860lowastlowastlowast (37052) mdash mdash mdash mdash mdash mdash mdash

Online socialnetwork mdash 0744lowastlowastlowast (24673) mdash minus0517lowastlowastlowast (minus12194) mdash mdash mdash mdash

Offline socialnetwork mdash mdash 0905lowastlowastlowast (45545) 1381lowastlowastlowast (32455) mdash mdash mdash mdash

Instrumentalnetwork mdash mdash mdash mdash 0735lowastlowastlowast (23892) mdash mdash 0016 (0818)

Emotionalnetwork mdash mdash mdash mdash mdash 0964lowastlowastlowast (81163) mdash 0793lowastlowastlowast (38244)

Mixednetwork mdash mdash mdash mdash mdash mdash 0880lowastlowastlowast (40217) 0189lowastlowastlowast (7352)

Adj-R2 0737 0554 0816 0858 0538 0931 0767 0942

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

6 Complexity

method was used to test the mediating effect If the 95confidence interval does not contain a zero then the me-diating effect becomes significant +e empirical regressionresults and significance test results for the mediating effectare shown in Table 3

+e results in Table 3 show that R2 in the three equationswas bigger than 07 and the goodness of fit for the model wasgood In equation (11) social network has a significantpositive effect on adolescent sports behaviorc 0860 plt 0001 In equation (12) social network has asignificant effect on social efficacy c 0839 plt 0001 Inequation (13) social network and social efficacy have asignificant effect on adolescent sports behaviorcprime 0122 b 0879 plt 0001 According to the mediatingeffect test procedure of Wen and Ye [18] the mediatingeffect of social efficacy is significant because a b c and cprime areall significant Further the Bootstrap method was used totest the mediating effect of social efficacy 5000 Bootstrapsamples were randomly selected from the original sample forindirect effect estimation Table 3 shows the estimated valueof the indirect effect of social efficacy to be 07374 and 95confidence intervals were [07007 07704] excluding zero+us the mediating effect of social efficacy becamesignificant

323 Moderating Effect of Self-Presentation As for themoderating effect of self-presentation the followingeconometric model was constructed

Sporti αSNi + βSPi + cSNi lowast SPi + PQ + e5 (14)

where SP is the self-presentation +e stepwise regressionmethod was used to estimate the moderating effect of self-presentation +e results are shown in Table 4

+e results of the study show that although both socialnetworking and self-presentation have significant positiveeffects on adolescent sports behavior the interaction termsof social network and self-presentation do not have anysignificant influence on adolescent sports behavior therebyindicating that the moderating effect of self-presentation isnot significant

In addition this study found that age has a negativepredictive effect on adolescent sports behavior that is theolder the adolescents are the less likely they are to exerciseGender had no significant effect on the prediction of ado-lescent sports behavior

33 Discussion

331 Influence of Social Networks on Adolescent SportsBehavior +e results of this study indicate that socialnetworks have significant influences on adolescent sportsbehavior demonstrating that adolescents with larger anddenser social networks have more frequent sports behaviorPeople are mutually linked to each other and so are theirsports behaviors +us the existence of social networksmeans that personsrsquo movements are interdependent [19]Sports behavior can be spread from person to person and

there is also a phenomenon called ldquopeer effectsrdquo [20] whichis just as the proverb says ldquoIf you live with a lame person youwill learn to limprdquo However it is interesting that with thestrengthening of the social network establishment the effectsof online social networks and offline social networks onadolescent sports behavior are not the same When ado-lescents have both an offline social network and an onlinesocial network the offline social network has a positive effecton adolescent sports while the online social network has anegative effect on adolescent sports +is outcome may bebecause at the present stage Chinese adolescents makefriends online mainly through games chats and other ways[21] Moreover adolescents in junior and senior highschools have heavy daily learning tasks and usually surf theInternet for leisure and entertainment +erefore mostadolescentsrsquo online social network relationships are enter-tainment groups Besides during the formation of onlinenetwork relationships most adolescents will use computersor mobile phones for a long time a factor that further re-duces the possibility of adolescent sports [22] Under theinfluence of the characteristics of network groups and thetime used for online communication offline social networkshave a negative influence on adolescent sports behavior

For social networks with different characteristics the in-strumental network emotional network and mixed networkhave positive effects on adolescent sports respectively In otherwords without the influence of others social relations such asthose with counselors teachers family members and friendshave positive effects on adolescent sports behaviors Howeverwhen adolescents have the above three kinds of social relationsat the same time the influence of the instrumental networkthat is counselors or teachers on adolescent sports behavior isnot significant +at is to say the influence of counselors orteachers on adolescent sports behavior is easily affected byother people +is outcome may be because the instrumentalnetwork connectivity is not strong [23] Even for teachers andstudents who live and study very closely there is no similarinfluence like those of friends indicating that although physicalactivity has the interpersonal infectivity of social network itsinfectivity changes based on relative activities and the gendersof friends [24] For example in a given social network lessactive runners tend to affect more active runners while theopposite is not true Both men and women affect men whileonly women affect other women [25]+e reason why it is easyfor physical activity behaviors to spread through adolescentfriendship or a family network is that first of all adolescents aremore willing to experience peer relations during adolescenceand their peer groups are based on common behaviors (sportscomputer games video games diet etc) +ese directly orindirectly affect their physical activities and health [26] Sec-ondly family relationship is the network relationship with thestrongest connection among adolescents Under the influenceof that strong connection an emotional network will have anobvious influence on adolescent sports behavior

Overall different types of social networks have positiveeffects on adolescent sports behavior separately but differentsocial networks are also interacted so their influences onadolescent sports behavior will change by the interactions

Complexity 7

+at is to say the effect of social networks on sports behaviornot only is influenced by the strength and density of socialnetworks but also influenced the interaction of differentsocial networks

332 Mediating Effect of Social Efficacy and the ModeratingEffect of Self-Presentation Social efficacy plays a mediatingrole in the relationship between social networking andadolescent sports behavior and these results can beexplained using the following perspectives

First according to the theoretical model of ldquotwo pathsand three layersrdquo of trust behavior social relationship is asituational factor and its influence on adolescent sportsbehavior is affected by internal cognition [27] In a positiveand extroverted social relationship adolescent sports be-haviors tend to receive more social support which then canenable adolescents to form a more positive internal cogni-tion of themselves and others thereby producing a strongermotivation to undertake sports [28]

Secondly from the perspective of self-efficacy theorysituational condition is one of the important factors that canaffect individual self-efficacy [29] If one is in a familiar andpleasant environment individualrsquos self-efficacy will increaseDoing sports is an important way to relieve adolescentpsychological stress and improve physical health +usadolescents receive positive demonstrations and responsesfrom their social network and their motivation to exercisewill strengthen in order to achieve the recognition of theirsocial relationships

+e moderating effect of self-presentation in the rela-tionship between a social network and adolescent sportsbehavior is not significant in that the positive effect of social

network on adolescent sports behavior is not affected byadolescentsrsquo self-presentation+is outcomemay be becauseon the one hand the level of self-presentation in Chineseadolescentsrsquo sports is low According to the Research Reporton Internet Usersrsquo Information Security released by the ChinaInternet Network Information Center (CNNIC 2018) morethan 73 of adolescentsrsquo self-presentation focuses on en-tertainment and study not so much on sports [30] On theother hand although self-presentation can reduce theloneliness of individuals and show them the enthusiasm ofsports behaviors it also presents certain negative informa-tion about sports such as fatigue and spending time therebyweakening the positive effect of self-presentation [31]

To sum up then this study concludes that social net-working not only directly predicts adolescent sports be-havior but also indirectly affects their actual sports behaviorthrough social efficacy +e implication of this mediatormodel is that adolescent sports behavior can also be gen-erated through interpersonal interaction and that behaviorcan be affected by individual cognition thereby showing therelevance of self-cognition gained by social interaction [32]

4 Conclusions

+e main conclusions of this study are the following

(1) +e social networks of adolescents have obviouspositive effects on their sports behavior Single onlinesocial networks and offline social networks instru-mental networks emotional networks and mixednetworks have obvious positive effects on adolescentsports behavior respectively However under theinfluence of multiple types of social networks offline

Table 4 Estimated results of moderating effect of self-presentation

Variables Adolescent sports behaviorConstant 0383lowast (1654) 0330lowast (1768) 0317lowast (1691)Age minus0028lowast (minus1764) minus0024lowast (minus1936) minus0024lowast (minus1916)Gender 0036 (0673) 0049 (1134) 0046 (1069)Self- presentation 0811lowastlowastlowast (30426) 0332lowastlowastlowast (9093) 0329lowastlowastlowast (8993)Social network mdash 0592lowastlowastlowast (16231) 0589lowastlowastlowast (16109)Social networklowastself-presentation mdash mdash 0014 (0891)Adj-R2 0654 0774 0774Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

Table 3 Estimated results and significance test results of mediating effect of social efficacy

Variables Equation (11)Sports

Equation (12)Social efficacy

Equation (13)Sports

Constant 0316 (1568) 0460lowastlowast (2148) minus0088 (minus1191)Age minus0023lowast (minus1667) minus0032lowastlowast (minus2228) 0006 (1123)Gender 0028 (0599) 0018 (0364) 0012 (0716)Social network 0860lowastlowastlowast (37052) 0839lowastlowastlowast (34099) 0122lowastlowastlowast (7916)Social efficacy mdash mdash 0879lowastlowastlowast (5691)Adj-R2 0737 0706 0956Indirect effect(s) of SN on sport

Intervening variable social contact Effect Boot standard error Boot LLCI Boot ULCI07374 00178 07007 07704

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

8 Complexity

social networking has a negative predictive effect onadolescent sports behavior while the effect of mixednetworks is not so obvious

(2) Social efficacy plays an intermediary role in the re-lationship between social networking and adolescentsports behavior +e moderating effect of self-pre-sentation is not significant

Although this paper found the importance of socialnetworks in adolescents sports behavior and discovered therole of social efficacy there are some limitations to this studywhich need to be addressed in future studies For example allvariables in this study were investigated using a questionnaireit canmeasure variables in static level butmay ignore diversityand complexity of variables furthermore this study foundthat the moderating effect of self-presentation is not signif-icant+is finding may have a theoretical rationale but it mayalso be due to the low degree of self-presentation of thesample In the future that sample needs to be expanded forfurther research and more precise evaluation

Data Availability

+e survey data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

+is study was supported by the MOE (Ministry of Edu-cation in China) Project of Humanities and Social Sciences(Project no 18YJC890012)

References

[1] L P Spear ldquo+e adolescent brain and age-related behavioralmanifestationsrdquo Neuroscience amp Biobehavioral Reviewsvol 24 no 4 pp 417ndash463 2000

[2] L Steinberg and K C Monahan ldquoAge differences in resis-tance to peer influencerdquo Developmental Psychology vol 43no 6 pp 1531ndash1543 2009

[3] S Sarah-Jeanne N R James C B Julie et al ldquoEffect of peersand friends on youth physical activity and motivation to bephysically activerdquo Journal of Pediatric Psychology vol 34no 2 pp 217ndash225 2009

[4] M Jennifer D L H Kayla L M Barnett et al ldquoFriendshipnetwork characteristics are associated with physical activityand sedentary behavior in early adolescencerdquo PLoS ONEvol 10 no 12 pp 1ndash15 2015

[5] M J Edwards R Jago J S Simon et al ldquo+e influence offriends and siblings on the physical activity and screenviewing behaviours of children aged 5-6 years a qualitativeanalysis of parent interviewsrdquo BMJ Open vol 5 no 5 ArticleID e006593 2015

[6] K D L Haye G Robins PMohr et al ldquoHow physical activityshapes and is shaped by adolescent friendshipsrdquo SocialScience amp Medicine vol 73 no 5 pp 719ndash728 2011

[7] Z Li F Xiong X Wang et al ldquoTopological influence-awarerecommendation on social networksrdquo Complexity vol 2019Article ID 6325654 12 pages 2019

[8] B Yu and X Hu ldquoToward training and assessing reproducibledata analysis in data science educationrdquo Data Intelligencevol 1 no 4 pp 333ndash344 2019

[9] A Bandura ldquoSelf-efficacy mechanism in human agencyrdquoAmerican Psychologist vol 37 no 2 pp 122ndash147 1982

[10] A Bandura Self-efficacy He Exercise of Control WorthPublishers New Nork NY USA 1997

[11] C Fan and A S Mak ldquoMeasuring social self-efficacy in aculturally diverse student populationrdquo Social Behavior andPersonality An International Journal vol 26 no 2 pp 131ndash144 1998

[12] M Wei D W Russell and R A Zakalik ldquoAdult attachmentsocial self-efficacy self-disclosure loneliness and subsequentdepression for freshman college students a longitudinal studyrdquoJournal of Counseling Psychology vol 52 no 4 pp 602ndash614 2005

[13] F Xiong W Shen H Chen et al ldquoExploiting implicit in-fluence from information propagation for social recom-mendationrdquo IEEE Transactions on Cybernetics 2019

[14] H Y Mao H C Hsu and S D Lee ldquoGender differences inrelated influential factors of regular exercise behavior amongpeople in Taiwan in 2007 a cross-sectional studyrdquo PLoS onevol 15 no 1 pp 1ndash22 2020

[15] S Park J Y Kang and L A Chadiha ldquoSocial network typeshealth and health-Care use among South Korean olderadultsrdquo Research on Aging vol 40 no 2 pp 131ndash154 2018

[16] E J Jeong and D H Kim ldquoSocial activities self-efficacy gameattitudes and game addictionrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 4 pp 213ndash221 2011

[17] J Kim and J-E R Lee ldquo+e Facebook paths to happinesseffects of the number of Facebook friends and self-presen-tation on subjective well-beingrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 6 pp 359ndash364 2011

[18] Z Wen and B Ye ldquoAnalyses of mediating effects the de-velopment of methods and modelsrdquo Advances in Psycho-logical Science vol 22 no 5 pp 731ndash745 2014

[19] L G Smith L Banting R Eime et al ldquo+e association be-tween social support and physical activity in older adultsasystematic reviewrdquo International Journal of Behavioral Nu-trition amp Physical Activity vol 14 no 1 p 56 2017

[20] K P Smith and N A Christakis ldquoSocial networks andhealthrdquo Annual Review of Sociology vol 34 no 1pp 405ndash429 2008

[21] S Aral and C Nicolaides ldquoExercise contagion in a globalsocial networkrdquo Nature Communications vol 8 Article ID14753 2017

[22] S Ji S Pan E Cambria et al ldquoA survey on knowledge graphsrepresentation acquisition and applicationsrdquo 2020 httparxivorgabs200200388

[23] T W Valente K Fujimoto C-P Chou and D Spruijt-MetzldquoAdolescent affiliations and adiposity a social networkanalysis of friendships and obesityrdquo Journal of AdolescentHealth vol 45 no 2 pp 202ndash204 2009

[24] P Manasatchakun P Chotiga J Hochwalder A RoxbergM Sandborgh and M Asp ldquoFactors associated with healthyaging among older persons in northeastern+ailandrdquo Journalof Cross-Cultural Gerontology vol 31 no 4 pp 369ndash3842016

[25] L Adam and P S Damon ldquo+e intersection of sport man-agement and sociology of sport research a social networkperspectiverdquo Sport Management Review vol 15 no 2pp 244ndash256 2012

Complexity 9

method was used to test the mediating effect If the 95confidence interval does not contain a zero then the me-diating effect becomes significant +e empirical regressionresults and significance test results for the mediating effectare shown in Table 3

+e results in Table 3 show that R2 in the three equationswas bigger than 07 and the goodness of fit for the model wasgood In equation (11) social network has a significantpositive effect on adolescent sports behaviorc 0860 plt 0001 In equation (12) social network has asignificant effect on social efficacy c 0839 plt 0001 Inequation (13) social network and social efficacy have asignificant effect on adolescent sports behaviorcprime 0122 b 0879 plt 0001 According to the mediatingeffect test procedure of Wen and Ye [18] the mediatingeffect of social efficacy is significant because a b c and cprime areall significant Further the Bootstrap method was used totest the mediating effect of social efficacy 5000 Bootstrapsamples were randomly selected from the original sample forindirect effect estimation Table 3 shows the estimated valueof the indirect effect of social efficacy to be 07374 and 95confidence intervals were [07007 07704] excluding zero+us the mediating effect of social efficacy becamesignificant

323 Moderating Effect of Self-Presentation As for themoderating effect of self-presentation the followingeconometric model was constructed

Sporti αSNi + βSPi + cSNi lowast SPi + PQ + e5 (14)

where SP is the self-presentation +e stepwise regressionmethod was used to estimate the moderating effect of self-presentation +e results are shown in Table 4

+e results of the study show that although both socialnetworking and self-presentation have significant positiveeffects on adolescent sports behavior the interaction termsof social network and self-presentation do not have anysignificant influence on adolescent sports behavior therebyindicating that the moderating effect of self-presentation isnot significant

In addition this study found that age has a negativepredictive effect on adolescent sports behavior that is theolder the adolescents are the less likely they are to exerciseGender had no significant effect on the prediction of ado-lescent sports behavior

33 Discussion

331 Influence of Social Networks on Adolescent SportsBehavior +e results of this study indicate that socialnetworks have significant influences on adolescent sportsbehavior demonstrating that adolescents with larger anddenser social networks have more frequent sports behaviorPeople are mutually linked to each other and so are theirsports behaviors +us the existence of social networksmeans that personsrsquo movements are interdependent [19]Sports behavior can be spread from person to person and

there is also a phenomenon called ldquopeer effectsrdquo [20] whichis just as the proverb says ldquoIf you live with a lame person youwill learn to limprdquo However it is interesting that with thestrengthening of the social network establishment the effectsof online social networks and offline social networks onadolescent sports behavior are not the same When ado-lescents have both an offline social network and an onlinesocial network the offline social network has a positive effecton adolescent sports while the online social network has anegative effect on adolescent sports +is outcome may bebecause at the present stage Chinese adolescents makefriends online mainly through games chats and other ways[21] Moreover adolescents in junior and senior highschools have heavy daily learning tasks and usually surf theInternet for leisure and entertainment +erefore mostadolescentsrsquo online social network relationships are enter-tainment groups Besides during the formation of onlinenetwork relationships most adolescents will use computersor mobile phones for a long time a factor that further re-duces the possibility of adolescent sports [22] Under theinfluence of the characteristics of network groups and thetime used for online communication offline social networkshave a negative influence on adolescent sports behavior

For social networks with different characteristics the in-strumental network emotional network and mixed networkhave positive effects on adolescent sports respectively In otherwords without the influence of others social relations such asthose with counselors teachers family members and friendshave positive effects on adolescent sports behaviors Howeverwhen adolescents have the above three kinds of social relationsat the same time the influence of the instrumental networkthat is counselors or teachers on adolescent sports behavior isnot significant +at is to say the influence of counselors orteachers on adolescent sports behavior is easily affected byother people +is outcome may be because the instrumentalnetwork connectivity is not strong [23] Even for teachers andstudents who live and study very closely there is no similarinfluence like those of friends indicating that although physicalactivity has the interpersonal infectivity of social network itsinfectivity changes based on relative activities and the gendersof friends [24] For example in a given social network lessactive runners tend to affect more active runners while theopposite is not true Both men and women affect men whileonly women affect other women [25]+e reason why it is easyfor physical activity behaviors to spread through adolescentfriendship or a family network is that first of all adolescents aremore willing to experience peer relations during adolescenceand their peer groups are based on common behaviors (sportscomputer games video games diet etc) +ese directly orindirectly affect their physical activities and health [26] Sec-ondly family relationship is the network relationship with thestrongest connection among adolescents Under the influenceof that strong connection an emotional network will have anobvious influence on adolescent sports behavior

Overall different types of social networks have positiveeffects on adolescent sports behavior separately but differentsocial networks are also interacted so their influences onadolescent sports behavior will change by the interactions

Complexity 7

+at is to say the effect of social networks on sports behaviornot only is influenced by the strength and density of socialnetworks but also influenced the interaction of differentsocial networks

332 Mediating Effect of Social Efficacy and the ModeratingEffect of Self-Presentation Social efficacy plays a mediatingrole in the relationship between social networking andadolescent sports behavior and these results can beexplained using the following perspectives

First according to the theoretical model of ldquotwo pathsand three layersrdquo of trust behavior social relationship is asituational factor and its influence on adolescent sportsbehavior is affected by internal cognition [27] In a positiveand extroverted social relationship adolescent sports be-haviors tend to receive more social support which then canenable adolescents to form a more positive internal cogni-tion of themselves and others thereby producing a strongermotivation to undertake sports [28]

Secondly from the perspective of self-efficacy theorysituational condition is one of the important factors that canaffect individual self-efficacy [29] If one is in a familiar andpleasant environment individualrsquos self-efficacy will increaseDoing sports is an important way to relieve adolescentpsychological stress and improve physical health +usadolescents receive positive demonstrations and responsesfrom their social network and their motivation to exercisewill strengthen in order to achieve the recognition of theirsocial relationships

+e moderating effect of self-presentation in the rela-tionship between a social network and adolescent sportsbehavior is not significant in that the positive effect of social

network on adolescent sports behavior is not affected byadolescentsrsquo self-presentation+is outcomemay be becauseon the one hand the level of self-presentation in Chineseadolescentsrsquo sports is low According to the Research Reporton Internet Usersrsquo Information Security released by the ChinaInternet Network Information Center (CNNIC 2018) morethan 73 of adolescentsrsquo self-presentation focuses on en-tertainment and study not so much on sports [30] On theother hand although self-presentation can reduce theloneliness of individuals and show them the enthusiasm ofsports behaviors it also presents certain negative informa-tion about sports such as fatigue and spending time therebyweakening the positive effect of self-presentation [31]

To sum up then this study concludes that social net-working not only directly predicts adolescent sports be-havior but also indirectly affects their actual sports behaviorthrough social efficacy +e implication of this mediatormodel is that adolescent sports behavior can also be gen-erated through interpersonal interaction and that behaviorcan be affected by individual cognition thereby showing therelevance of self-cognition gained by social interaction [32]

4 Conclusions

+e main conclusions of this study are the following

(1) +e social networks of adolescents have obviouspositive effects on their sports behavior Single onlinesocial networks and offline social networks instru-mental networks emotional networks and mixednetworks have obvious positive effects on adolescentsports behavior respectively However under theinfluence of multiple types of social networks offline

Table 4 Estimated results of moderating effect of self-presentation

Variables Adolescent sports behaviorConstant 0383lowast (1654) 0330lowast (1768) 0317lowast (1691)Age minus0028lowast (minus1764) minus0024lowast (minus1936) minus0024lowast (minus1916)Gender 0036 (0673) 0049 (1134) 0046 (1069)Self- presentation 0811lowastlowastlowast (30426) 0332lowastlowastlowast (9093) 0329lowastlowastlowast (8993)Social network mdash 0592lowastlowastlowast (16231) 0589lowastlowastlowast (16109)Social networklowastself-presentation mdash mdash 0014 (0891)Adj-R2 0654 0774 0774Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

Table 3 Estimated results and significance test results of mediating effect of social efficacy

Variables Equation (11)Sports

Equation (12)Social efficacy

Equation (13)Sports

Constant 0316 (1568) 0460lowastlowast (2148) minus0088 (minus1191)Age minus0023lowast (minus1667) minus0032lowastlowast (minus2228) 0006 (1123)Gender 0028 (0599) 0018 (0364) 0012 (0716)Social network 0860lowastlowastlowast (37052) 0839lowastlowastlowast (34099) 0122lowastlowastlowast (7916)Social efficacy mdash mdash 0879lowastlowastlowast (5691)Adj-R2 0737 0706 0956Indirect effect(s) of SN on sport

Intervening variable social contact Effect Boot standard error Boot LLCI Boot ULCI07374 00178 07007 07704

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

8 Complexity

social networking has a negative predictive effect onadolescent sports behavior while the effect of mixednetworks is not so obvious

(2) Social efficacy plays an intermediary role in the re-lationship between social networking and adolescentsports behavior +e moderating effect of self-pre-sentation is not significant

Although this paper found the importance of socialnetworks in adolescents sports behavior and discovered therole of social efficacy there are some limitations to this studywhich need to be addressed in future studies For example allvariables in this study were investigated using a questionnaireit canmeasure variables in static level butmay ignore diversityand complexity of variables furthermore this study foundthat the moderating effect of self-presentation is not signif-icant+is finding may have a theoretical rationale but it mayalso be due to the low degree of self-presentation of thesample In the future that sample needs to be expanded forfurther research and more precise evaluation

Data Availability

+e survey data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

+is study was supported by the MOE (Ministry of Edu-cation in China) Project of Humanities and Social Sciences(Project no 18YJC890012)

References

[1] L P Spear ldquo+e adolescent brain and age-related behavioralmanifestationsrdquo Neuroscience amp Biobehavioral Reviewsvol 24 no 4 pp 417ndash463 2000

[2] L Steinberg and K C Monahan ldquoAge differences in resis-tance to peer influencerdquo Developmental Psychology vol 43no 6 pp 1531ndash1543 2009

[3] S Sarah-Jeanne N R James C B Julie et al ldquoEffect of peersand friends on youth physical activity and motivation to bephysically activerdquo Journal of Pediatric Psychology vol 34no 2 pp 217ndash225 2009

[4] M Jennifer D L H Kayla L M Barnett et al ldquoFriendshipnetwork characteristics are associated with physical activityand sedentary behavior in early adolescencerdquo PLoS ONEvol 10 no 12 pp 1ndash15 2015

[5] M J Edwards R Jago J S Simon et al ldquo+e influence offriends and siblings on the physical activity and screenviewing behaviours of children aged 5-6 years a qualitativeanalysis of parent interviewsrdquo BMJ Open vol 5 no 5 ArticleID e006593 2015

[6] K D L Haye G Robins PMohr et al ldquoHow physical activityshapes and is shaped by adolescent friendshipsrdquo SocialScience amp Medicine vol 73 no 5 pp 719ndash728 2011

[7] Z Li F Xiong X Wang et al ldquoTopological influence-awarerecommendation on social networksrdquo Complexity vol 2019Article ID 6325654 12 pages 2019

[8] B Yu and X Hu ldquoToward training and assessing reproducibledata analysis in data science educationrdquo Data Intelligencevol 1 no 4 pp 333ndash344 2019

[9] A Bandura ldquoSelf-efficacy mechanism in human agencyrdquoAmerican Psychologist vol 37 no 2 pp 122ndash147 1982

[10] A Bandura Self-efficacy He Exercise of Control WorthPublishers New Nork NY USA 1997

[11] C Fan and A S Mak ldquoMeasuring social self-efficacy in aculturally diverse student populationrdquo Social Behavior andPersonality An International Journal vol 26 no 2 pp 131ndash144 1998

[12] M Wei D W Russell and R A Zakalik ldquoAdult attachmentsocial self-efficacy self-disclosure loneliness and subsequentdepression for freshman college students a longitudinal studyrdquoJournal of Counseling Psychology vol 52 no 4 pp 602ndash614 2005

[13] F Xiong W Shen H Chen et al ldquoExploiting implicit in-fluence from information propagation for social recom-mendationrdquo IEEE Transactions on Cybernetics 2019

[14] H Y Mao H C Hsu and S D Lee ldquoGender differences inrelated influential factors of regular exercise behavior amongpeople in Taiwan in 2007 a cross-sectional studyrdquo PLoS onevol 15 no 1 pp 1ndash22 2020

[15] S Park J Y Kang and L A Chadiha ldquoSocial network typeshealth and health-Care use among South Korean olderadultsrdquo Research on Aging vol 40 no 2 pp 131ndash154 2018

[16] E J Jeong and D H Kim ldquoSocial activities self-efficacy gameattitudes and game addictionrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 4 pp 213ndash221 2011

[17] J Kim and J-E R Lee ldquo+e Facebook paths to happinesseffects of the number of Facebook friends and self-presen-tation on subjective well-beingrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 6 pp 359ndash364 2011

[18] Z Wen and B Ye ldquoAnalyses of mediating effects the de-velopment of methods and modelsrdquo Advances in Psycho-logical Science vol 22 no 5 pp 731ndash745 2014

[19] L G Smith L Banting R Eime et al ldquo+e association be-tween social support and physical activity in older adultsasystematic reviewrdquo International Journal of Behavioral Nu-trition amp Physical Activity vol 14 no 1 p 56 2017

[20] K P Smith and N A Christakis ldquoSocial networks andhealthrdquo Annual Review of Sociology vol 34 no 1pp 405ndash429 2008

[21] S Aral and C Nicolaides ldquoExercise contagion in a globalsocial networkrdquo Nature Communications vol 8 Article ID14753 2017

[22] S Ji S Pan E Cambria et al ldquoA survey on knowledge graphsrepresentation acquisition and applicationsrdquo 2020 httparxivorgabs200200388

[23] T W Valente K Fujimoto C-P Chou and D Spruijt-MetzldquoAdolescent affiliations and adiposity a social networkanalysis of friendships and obesityrdquo Journal of AdolescentHealth vol 45 no 2 pp 202ndash204 2009

[24] P Manasatchakun P Chotiga J Hochwalder A RoxbergM Sandborgh and M Asp ldquoFactors associated with healthyaging among older persons in northeastern+ailandrdquo Journalof Cross-Cultural Gerontology vol 31 no 4 pp 369ndash3842016

[25] L Adam and P S Damon ldquo+e intersection of sport man-agement and sociology of sport research a social networkperspectiverdquo Sport Management Review vol 15 no 2pp 244ndash256 2012

Complexity 9

+at is to say the effect of social networks on sports behaviornot only is influenced by the strength and density of socialnetworks but also influenced the interaction of differentsocial networks

332 Mediating Effect of Social Efficacy and the ModeratingEffect of Self-Presentation Social efficacy plays a mediatingrole in the relationship between social networking andadolescent sports behavior and these results can beexplained using the following perspectives

First according to the theoretical model of ldquotwo pathsand three layersrdquo of trust behavior social relationship is asituational factor and its influence on adolescent sportsbehavior is affected by internal cognition [27] In a positiveand extroverted social relationship adolescent sports be-haviors tend to receive more social support which then canenable adolescents to form a more positive internal cogni-tion of themselves and others thereby producing a strongermotivation to undertake sports [28]

Secondly from the perspective of self-efficacy theorysituational condition is one of the important factors that canaffect individual self-efficacy [29] If one is in a familiar andpleasant environment individualrsquos self-efficacy will increaseDoing sports is an important way to relieve adolescentpsychological stress and improve physical health +usadolescents receive positive demonstrations and responsesfrom their social network and their motivation to exercisewill strengthen in order to achieve the recognition of theirsocial relationships

+e moderating effect of self-presentation in the rela-tionship between a social network and adolescent sportsbehavior is not significant in that the positive effect of social

network on adolescent sports behavior is not affected byadolescentsrsquo self-presentation+is outcomemay be becauseon the one hand the level of self-presentation in Chineseadolescentsrsquo sports is low According to the Research Reporton Internet Usersrsquo Information Security released by the ChinaInternet Network Information Center (CNNIC 2018) morethan 73 of adolescentsrsquo self-presentation focuses on en-tertainment and study not so much on sports [30] On theother hand although self-presentation can reduce theloneliness of individuals and show them the enthusiasm ofsports behaviors it also presents certain negative informa-tion about sports such as fatigue and spending time therebyweakening the positive effect of self-presentation [31]

To sum up then this study concludes that social net-working not only directly predicts adolescent sports be-havior but also indirectly affects their actual sports behaviorthrough social efficacy +e implication of this mediatormodel is that adolescent sports behavior can also be gen-erated through interpersonal interaction and that behaviorcan be affected by individual cognition thereby showing therelevance of self-cognition gained by social interaction [32]

4 Conclusions

+e main conclusions of this study are the following

(1) +e social networks of adolescents have obviouspositive effects on their sports behavior Single onlinesocial networks and offline social networks instru-mental networks emotional networks and mixednetworks have obvious positive effects on adolescentsports behavior respectively However under theinfluence of multiple types of social networks offline

Table 4 Estimated results of moderating effect of self-presentation

Variables Adolescent sports behaviorConstant 0383lowast (1654) 0330lowast (1768) 0317lowast (1691)Age minus0028lowast (minus1764) minus0024lowast (minus1936) minus0024lowast (minus1916)Gender 0036 (0673) 0049 (1134) 0046 (1069)Self- presentation 0811lowastlowastlowast (30426) 0332lowastlowastlowast (9093) 0329lowastlowastlowast (8993)Social network mdash 0592lowastlowastlowast (16231) 0589lowastlowastlowast (16109)Social networklowastself-presentation mdash mdash 0014 (0891)Adj-R2 0654 0774 0774Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

Table 3 Estimated results and significance test results of mediating effect of social efficacy

Variables Equation (11)Sports

Equation (12)Social efficacy

Equation (13)Sports

Constant 0316 (1568) 0460lowastlowast (2148) minus0088 (minus1191)Age minus0023lowast (minus1667) minus0032lowastlowast (minus2228) 0006 (1123)Gender 0028 (0599) 0018 (0364) 0012 (0716)Social network 0860lowastlowastlowast (37052) 0839lowastlowastlowast (34099) 0122lowastlowastlowast (7916)Social efficacy mdash mdash 0879lowastlowastlowast (5691)Adj-R2 0737 0706 0956Indirect effect(s) of SN on sport

Intervening variable social contact Effect Boot standard error Boot LLCI Boot ULCI07374 00178 07007 07704

Note t value is in parentheses Symbols lowast lowastlowast and lowastlowastlowast indicate that coefficient is significant at 10 5 and 1 levels

8 Complexity

social networking has a negative predictive effect onadolescent sports behavior while the effect of mixednetworks is not so obvious

(2) Social efficacy plays an intermediary role in the re-lationship between social networking and adolescentsports behavior +e moderating effect of self-pre-sentation is not significant

Although this paper found the importance of socialnetworks in adolescents sports behavior and discovered therole of social efficacy there are some limitations to this studywhich need to be addressed in future studies For example allvariables in this study were investigated using a questionnaireit canmeasure variables in static level butmay ignore diversityand complexity of variables furthermore this study foundthat the moderating effect of self-presentation is not signif-icant+is finding may have a theoretical rationale but it mayalso be due to the low degree of self-presentation of thesample In the future that sample needs to be expanded forfurther research and more precise evaluation

Data Availability

+e survey data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

+is study was supported by the MOE (Ministry of Edu-cation in China) Project of Humanities and Social Sciences(Project no 18YJC890012)

References

[1] L P Spear ldquo+e adolescent brain and age-related behavioralmanifestationsrdquo Neuroscience amp Biobehavioral Reviewsvol 24 no 4 pp 417ndash463 2000

[2] L Steinberg and K C Monahan ldquoAge differences in resis-tance to peer influencerdquo Developmental Psychology vol 43no 6 pp 1531ndash1543 2009

[3] S Sarah-Jeanne N R James C B Julie et al ldquoEffect of peersand friends on youth physical activity and motivation to bephysically activerdquo Journal of Pediatric Psychology vol 34no 2 pp 217ndash225 2009

[4] M Jennifer D L H Kayla L M Barnett et al ldquoFriendshipnetwork characteristics are associated with physical activityand sedentary behavior in early adolescencerdquo PLoS ONEvol 10 no 12 pp 1ndash15 2015

[5] M J Edwards R Jago J S Simon et al ldquo+e influence offriends and siblings on the physical activity and screenviewing behaviours of children aged 5-6 years a qualitativeanalysis of parent interviewsrdquo BMJ Open vol 5 no 5 ArticleID e006593 2015

[6] K D L Haye G Robins PMohr et al ldquoHow physical activityshapes and is shaped by adolescent friendshipsrdquo SocialScience amp Medicine vol 73 no 5 pp 719ndash728 2011

[7] Z Li F Xiong X Wang et al ldquoTopological influence-awarerecommendation on social networksrdquo Complexity vol 2019Article ID 6325654 12 pages 2019

[8] B Yu and X Hu ldquoToward training and assessing reproducibledata analysis in data science educationrdquo Data Intelligencevol 1 no 4 pp 333ndash344 2019

[9] A Bandura ldquoSelf-efficacy mechanism in human agencyrdquoAmerican Psychologist vol 37 no 2 pp 122ndash147 1982

[10] A Bandura Self-efficacy He Exercise of Control WorthPublishers New Nork NY USA 1997

[11] C Fan and A S Mak ldquoMeasuring social self-efficacy in aculturally diverse student populationrdquo Social Behavior andPersonality An International Journal vol 26 no 2 pp 131ndash144 1998

[12] M Wei D W Russell and R A Zakalik ldquoAdult attachmentsocial self-efficacy self-disclosure loneliness and subsequentdepression for freshman college students a longitudinal studyrdquoJournal of Counseling Psychology vol 52 no 4 pp 602ndash614 2005

[13] F Xiong W Shen H Chen et al ldquoExploiting implicit in-fluence from information propagation for social recom-mendationrdquo IEEE Transactions on Cybernetics 2019

[14] H Y Mao H C Hsu and S D Lee ldquoGender differences inrelated influential factors of regular exercise behavior amongpeople in Taiwan in 2007 a cross-sectional studyrdquo PLoS onevol 15 no 1 pp 1ndash22 2020

[15] S Park J Y Kang and L A Chadiha ldquoSocial network typeshealth and health-Care use among South Korean olderadultsrdquo Research on Aging vol 40 no 2 pp 131ndash154 2018

[16] E J Jeong and D H Kim ldquoSocial activities self-efficacy gameattitudes and game addictionrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 4 pp 213ndash221 2011

[17] J Kim and J-E R Lee ldquo+e Facebook paths to happinesseffects of the number of Facebook friends and self-presen-tation on subjective well-beingrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 6 pp 359ndash364 2011

[18] Z Wen and B Ye ldquoAnalyses of mediating effects the de-velopment of methods and modelsrdquo Advances in Psycho-logical Science vol 22 no 5 pp 731ndash745 2014

[19] L G Smith L Banting R Eime et al ldquo+e association be-tween social support and physical activity in older adultsasystematic reviewrdquo International Journal of Behavioral Nu-trition amp Physical Activity vol 14 no 1 p 56 2017

[20] K P Smith and N A Christakis ldquoSocial networks andhealthrdquo Annual Review of Sociology vol 34 no 1pp 405ndash429 2008

[21] S Aral and C Nicolaides ldquoExercise contagion in a globalsocial networkrdquo Nature Communications vol 8 Article ID14753 2017

[22] S Ji S Pan E Cambria et al ldquoA survey on knowledge graphsrepresentation acquisition and applicationsrdquo 2020 httparxivorgabs200200388

[23] T W Valente K Fujimoto C-P Chou and D Spruijt-MetzldquoAdolescent affiliations and adiposity a social networkanalysis of friendships and obesityrdquo Journal of AdolescentHealth vol 45 no 2 pp 202ndash204 2009

[24] P Manasatchakun P Chotiga J Hochwalder A RoxbergM Sandborgh and M Asp ldquoFactors associated with healthyaging among older persons in northeastern+ailandrdquo Journalof Cross-Cultural Gerontology vol 31 no 4 pp 369ndash3842016

[25] L Adam and P S Damon ldquo+e intersection of sport man-agement and sociology of sport research a social networkperspectiverdquo Sport Management Review vol 15 no 2pp 244ndash256 2012

Complexity 9

social networking has a negative predictive effect onadolescent sports behavior while the effect of mixednetworks is not so obvious

(2) Social efficacy plays an intermediary role in the re-lationship between social networking and adolescentsports behavior +e moderating effect of self-pre-sentation is not significant

Although this paper found the importance of socialnetworks in adolescents sports behavior and discovered therole of social efficacy there are some limitations to this studywhich need to be addressed in future studies For example allvariables in this study were investigated using a questionnaireit canmeasure variables in static level butmay ignore diversityand complexity of variables furthermore this study foundthat the moderating effect of self-presentation is not signif-icant+is finding may have a theoretical rationale but it mayalso be due to the low degree of self-presentation of thesample In the future that sample needs to be expanded forfurther research and more precise evaluation

Data Availability

+e survey data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

+is study was supported by the MOE (Ministry of Edu-cation in China) Project of Humanities and Social Sciences(Project no 18YJC890012)

References

[1] L P Spear ldquo+e adolescent brain and age-related behavioralmanifestationsrdquo Neuroscience amp Biobehavioral Reviewsvol 24 no 4 pp 417ndash463 2000

[2] L Steinberg and K C Monahan ldquoAge differences in resis-tance to peer influencerdquo Developmental Psychology vol 43no 6 pp 1531ndash1543 2009

[3] S Sarah-Jeanne N R James C B Julie et al ldquoEffect of peersand friends on youth physical activity and motivation to bephysically activerdquo Journal of Pediatric Psychology vol 34no 2 pp 217ndash225 2009

[4] M Jennifer D L H Kayla L M Barnett et al ldquoFriendshipnetwork characteristics are associated with physical activityand sedentary behavior in early adolescencerdquo PLoS ONEvol 10 no 12 pp 1ndash15 2015

[5] M J Edwards R Jago J S Simon et al ldquo+e influence offriends and siblings on the physical activity and screenviewing behaviours of children aged 5-6 years a qualitativeanalysis of parent interviewsrdquo BMJ Open vol 5 no 5 ArticleID e006593 2015

[6] K D L Haye G Robins PMohr et al ldquoHow physical activityshapes and is shaped by adolescent friendshipsrdquo SocialScience amp Medicine vol 73 no 5 pp 719ndash728 2011

[7] Z Li F Xiong X Wang et al ldquoTopological influence-awarerecommendation on social networksrdquo Complexity vol 2019Article ID 6325654 12 pages 2019

[8] B Yu and X Hu ldquoToward training and assessing reproducibledata analysis in data science educationrdquo Data Intelligencevol 1 no 4 pp 333ndash344 2019

[9] A Bandura ldquoSelf-efficacy mechanism in human agencyrdquoAmerican Psychologist vol 37 no 2 pp 122ndash147 1982

[10] A Bandura Self-efficacy He Exercise of Control WorthPublishers New Nork NY USA 1997

[11] C Fan and A S Mak ldquoMeasuring social self-efficacy in aculturally diverse student populationrdquo Social Behavior andPersonality An International Journal vol 26 no 2 pp 131ndash144 1998

[12] M Wei D W Russell and R A Zakalik ldquoAdult attachmentsocial self-efficacy self-disclosure loneliness and subsequentdepression for freshman college students a longitudinal studyrdquoJournal of Counseling Psychology vol 52 no 4 pp 602ndash614 2005

[13] F Xiong W Shen H Chen et al ldquoExploiting implicit in-fluence from information propagation for social recom-mendationrdquo IEEE Transactions on Cybernetics 2019

[14] H Y Mao H C Hsu and S D Lee ldquoGender differences inrelated influential factors of regular exercise behavior amongpeople in Taiwan in 2007 a cross-sectional studyrdquo PLoS onevol 15 no 1 pp 1ndash22 2020

[15] S Park J Y Kang and L A Chadiha ldquoSocial network typeshealth and health-Care use among South Korean olderadultsrdquo Research on Aging vol 40 no 2 pp 131ndash154 2018

[16] E J Jeong and D H Kim ldquoSocial activities self-efficacy gameattitudes and game addictionrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 4 pp 213ndash221 2011

[17] J Kim and J-E R Lee ldquo+e Facebook paths to happinesseffects of the number of Facebook friends and self-presen-tation on subjective well-beingrdquo Cyberpsychology Behaviorand Social Networking vol 14 no 6 pp 359ndash364 2011

[18] Z Wen and B Ye ldquoAnalyses of mediating effects the de-velopment of methods and modelsrdquo Advances in Psycho-logical Science vol 22 no 5 pp 731ndash745 2014

[19] L G Smith L Banting R Eime et al ldquo+e association be-tween social support and physical activity in older adultsasystematic reviewrdquo International Journal of Behavioral Nu-trition amp Physical Activity vol 14 no 1 p 56 2017

[20] K P Smith and N A Christakis ldquoSocial networks andhealthrdquo Annual Review of Sociology vol 34 no 1pp 405ndash429 2008

[21] S Aral and C Nicolaides ldquoExercise contagion in a globalsocial networkrdquo Nature Communications vol 8 Article ID14753 2017

[22] S Ji S Pan E Cambria et al ldquoA survey on knowledge graphsrepresentation acquisition and applicationsrdquo 2020 httparxivorgabs200200388

[23] T W Valente K Fujimoto C-P Chou and D Spruijt-MetzldquoAdolescent affiliations and adiposity a social networkanalysis of friendships and obesityrdquo Journal of AdolescentHealth vol 45 no 2 pp 202ndash204 2009

[24] P Manasatchakun P Chotiga J Hochwalder A RoxbergM Sandborgh and M Asp ldquoFactors associated with healthyaging among older persons in northeastern+ailandrdquo Journalof Cross-Cultural Gerontology vol 31 no 4 pp 369ndash3842016

[25] L Adam and P S Damon ldquo+e intersection of sport man-agement and sociology of sport research a social networkperspectiverdquo Sport Management Review vol 15 no 2pp 244ndash256 2012

Complexity 9

[26] Z Wu S Pan F Chen et al ldquoA comprehensive survey ongraph neural networksrdquo IEEE Transactions on Neural Net-works and Learning Systems vol 99 2020

[27] L N Howard ldquoSport sociology NASSS and undergraduateeducation in the United States a social network perspectivefor developing the fieldrdquo Sociology of Sport Journal vol 27no 1 pp 76ndash88 2010

[28] Q Zhu ldquoSelf-disclosure in online support groups for peopleliving with depressionrdquo Masterrsquos +esis National Universityof Singapore Singapore 2011

[29] F Xiong X Wang S Pan H Yang H Wang and C ZhangldquoSocial recommendation with evolutionary opinion dynam-icsrdquo IEEE Transactions on Systems Man and CyberneticsSystems 2019

[30] L S Eller E L Lev C Yuan and A V Watkins ldquoDescribingself-care self-efficacy definition measurement outcomes andImplicationsdefinitionmeasurement outcomes and im-plicationsrdquo International Journal of Nursing Knowledgevol 29 no 1 pp 38ndash48 2018

[31] B Cornwell L P Schumm E O Laumann et al ldquoAssessmentof social network change in a national longitudinal surveyrdquoJournals of Gerontology vol 69 no 2 pp 75ndash82 2014

[32] L Guo Q Zhang W Hu Z Sun and Y Qu ldquoLearning tocomplete knowledge graphs with deep sequential modelsrdquoData Intelligence vol 1 no 3 pp 224ndash243 2019

10 Complexity