Journal Article Review Prosocial Behavior

J Youth Adolescence (2007) 36:255–264 DOI 10.1007/s10964-006-9085-y

O R I G I N A L PA P E R

The Contribution of Organized Youth Sport to Antisocial and Prosocial Behavior in Adolescent Athletes Esther A. Rutten · Geert Jan J. M. Stams · Gert J. J. Biesta · Carlo Schuengel · Evelien Dirks · Jan B. Hoeksma

Received: 1 February 2005 / Accepted: 18 April 2005 / Published online: 25 July 2006 C© Springer Science+Business Media, Inc. 2006

Abstract In this study, we investigated the contribution of organized youth sport to antisocial and prosocial behavior in adolescent athletes. The sample consisted of N = 260 male and female soccer players and competitive swimmers, 12 to 18 years of age. Multilevel regression analysis revealed that 8% of the variance in antisocial behavior and 7% of the variance in prosocial behavior could be attributed to charac- teristics of the sporting environment. Results suggested that coaches who maintain good relationships with their athletes reduce antisocial behavior, and that exposure to relatively high levels of sociomoral reasoning within the immediate context of sporting activities promotes prosocial behavior. These results point to specific aspects of adolescents’ par- ticipation in sport that can be used to realize the educational potential of organized youth sport.

Keywords Antisocial behavior . Prosocial behavior .

Moral reasoning . Coach-athlete relationship quality .

Sociomoral atmosphere

E. A. Rutten (�) She is currently writing her Ph.D. thesis on sports and education. Interests include moral development, behavioral adaptation, and delinquency. University of Amsterdam, Faculty of Social and Behavioral Sciences, Department of Education, P.O. Box 94208, 1090 GE Amsterdam, The Netherlands. e-mail: [email protected], [email protected]

G. J. J. M. Stams Associate Professor of Special Education at the University of Amsterdam, The Netherlands. His research interests include socio-emotional development and moral education.

G. J. J. Biesta Professor of Educational Theory in the School of Education and life long learning, University of Exeter, UK. His research focuses on the role of communication in education.

C. Schuengel Professor of Special Education at the Free University of Amsterdam, The Netherlands. His research interests concern socio-emotional development and developmental psychopathology, especially within the framework of attachment theory.

E. Dirks She is working on a Ph.D. thesis on dyslexia. Her interests include socio-emotional development and learning problems. Free University of Amsterdam, The Netherlands.

J. B. Hoeksma Assistant Professor at the Department of Developmental Psychology of Free University of Amsterdam, The Netherlands. Interests include methods of developmental research and multilevel modeling.

Introduction

Adolescents do not engage in sports in order to be educated. Yet, each social practice in which they participate could have an educational influence. Young athletes not only learn the skills and knowledge they need to perform their sports; they also learn sport-related rules and norms. Between 12 and 18 years of age, no fewer than 68% of Dutch adolescents participate in organized youth sport (CBS, 1999). It is a context in which adolescents voluntarily participate. Young athletes subject themselves to the authority of adults, and expose themselves to the influence of peers. Notably, the socializing influence of peers in leisure activities and prac- tices is substantial, and could even be greater in this domain than in the family and school context (Cotterell, 1996; Em- ler and Reicher, 1995; Mahoney, 2000; Mahoney and Stattin, 2000; Weiss and Smith, 2002). Thus it is important to study to what extent and how participation in organized youth sport exerts an influence on young people’s behavior.

Springer

256 J Youth Adolescence (2007) 36:255–264

Some research suggests that competitive and performance oriented social contexts actually promote antisocial behav- ior (Anderson and Morrow, 1995; Kohn, 1986; Stephens and Bredemeier, 1996) and have a negative effect on prosocial be- havior (Kleiber and Roberts, 1981). Shields and Bredemeier (1995) suggest that organized youth sport may promote anti- social behavior, as sport is a context of moral release, being based on competition, self-interest and a suspension of re- lational responsibility, while moral deliberation is reduced by formal and informal rules. The empirical evidence for an association between organized youth sport and juvenile an- tisocial behavior, however, remains equivocal (Begg et al., 1996; Coalter, 1989; Duncan et al., 2002; Mutrie and Parfitt, 1998; Robins, 1990; Vazsonyi et al., 2002). Duncan et al. (2002) assume that contradictory research outcomes may be due to the types of sport activity and behaviors examined. For instance, participation in contact sports has been associ- ated with higher perceived legitimacy of aggressive behavior (Conroy et al., 2001) and lower levels of moral functioning (Kavussanu and Ntoumanis, 2003) in comparison with non- contact sports. In addition, Vallerand et al. (1997) found that team sport athletes showed lower levels of concern for the opponent than individual sport athletes.

Arnold (1994, 2001), in contrast, argues that sport should be understood as a practice imbued with moral values. Sport not only entails respect for the rules of the game, but also relates to respect for one’s opponents, equal opportunities to perform well, mutual co-operation, fairness, and sportper- sonship. These elements are considered to be constitutive of sport, and may be summarized by the concept of “fair play” (Loland and McNamee, 2000). Van Bottenburg and Schuyt (1996) contend that sporting activity fosters the development of important virtues, such as team spirit and social respon- sibility. Coakley (1984) emphasized that sporting activity stimulates social-cognitive competences, such as role-taking ability, as athletes must constantly anticipate their opponents’ actions or moves. While role-taking ability is a necessary condition for growth in moral reasoning (Kohlberg et al., 1987), virtuous dispositions like self-control or perseverance could be important for the translation of moral reasoning into moral behavior (Matsuba and Walker, 1998). Finally, there is empirical evidence showing that school extracurricular activity participation (Mahoney, 2000) and supervised and structured leisure activities (Vazsonyi et al., 2002), includ- ing organized youth sport, are associated with less antisocial behavior in adolescents.

In the present study, we examine the contribution of orga- nized youth sport to antisocial and prosocial behavior among adolescent soccer players and competitive swimmers, focus- ing on educationally relevant factors, including sociomoral atmosphere of the sporting environment, sociomoral rea- soning about sport dilemmas, and coach-athlete relationship quality. It has been demonstrated that a favorable sociomoral

climate of the environment (e.g. Guivernau and Duda, 2002; Ommundsen et al., 2003; Power et al., 1989; Stephens, 2000), and mature sociomoral reasoning (Blasi, 1980; Bredemeier, 1994; Nelson et al., 1990; Ommundsen et al., 2003) are both related to less antisocial and more prosocial behavior. From studies on natural occurring mentoring (e.g. Darling et al., 2002; Rhodes, 2002; Zimmerman et al., 2002) it can be derived that high quality coach-athlete relationships may protect against antisocial behavior and promote proso- cial behavior, since the coach serves as a positive role model and provides emotional support. It is important to note that relational support from adults has been shown to be a strong predictor of behavioral adjustment in adolescents (DuBois et al., 2002; Scholte et al., 2001; Wills and Cleary, 1996).

In soccer and competitive swimming, individual athletes are always part of specific teams or exercise groups. Since the present study focuses on these types of sport, we can take advantage of the nested structure of the data using multilevel regression analysis (Goldstein, 1995), which permits the simultaneous analysis of the degree to which individual char- acteristics of athletes (individual level) and characteristics of teams or coaches (contextual level) are associated with antisocial and prosocial behavior. Whereas at the individual level sociomoral atmosphere reflects individual perceptions of the sociomoral climate, at the contextual level sociomoral atmosphere is a shared perspective by team members. So- ciomoral reasoning can be evaluated at the individual level, reflecting differences in moral reasoning among individual athletes, or at the contextual level, reflecting differences in moral reasoning between teams. Finally, at the individual level coach-athlete relationship quality refers to the athlete’s perception of his or her relationship with the coach, whereas at the contextual level coach-athlete relationship quality reflects the way a coach interacts with his athletes. As such, coach-athlete relationship quality can have an effect on antisocial and prosocial behavior at both levels. Eventually, antisocial and prosocial behavior will depend on characteristics of individual athletes and characteristics of teams or coaches, that is, the social context.

Hypotheses

We hypothesize that an advantageous sociomoral atmo- sphere of the sporting environment—characterized by mu- tual respect, care, trust, responsibility, and shared prosocial norms about what constitutes appropriate behavior—and a relatively high level of sociomoral reasoning about sport dilemmas, will be related to less antisocial behavior and more prosocial behavior in adolescent athletes. Next, we hypothesize that high quality coach-athlete relationships will be related to less antisocial behavior and more prosocial behavior in adolescent athletes.

Springer

J Youth Adolescence (2007) 36:255–264 257

Both hypotheses will be evaluated at the individual and contextual level, using multilevel regression analysis. It should be noted that interpretations of effects at the contex- tual level are most straightforward, as these effects are inde- pendent from the contribution the athlete makes to the team or the relationship with the coach. In cross-sectional mul- tilevel studies contextual effects strengthen the confidence in the causal interpretation of effects (Kreft and De Leeuw, 1998; Snijders and Bosker, 1999). Because self-selection can be a possible alternative explanation for what appear to be contextual effects, we will control for demographic variables that might play a role in the selection of a sport, a club, or a team, such as age, sex, level of education, current extent and years of sport participation, cultural background, and socioe- conomic status. Moreover, we will control for type of sport, as soccer and competitive swimming are different types of sport with respect to physical contact and individual or team performance.

Method

Participants

We included soccer and competitive swimming into our study, as these sports constitute a team and an individual sport, respectively, having high participation rates in The Netherlands. Both sports are performed by adolescents from lower and lower-middle-class socioeconomic backgrounds. Whereas soccer is a contact sport, competitive swimming represents an individual non-contact sport that is performed in exercise groups or teams. A total number of N = 10 sports clubs participated in our study: n = 6 soccer clubs (n = 187 adolescents) and n = 4 swimming clubs (n = 73 adolescents). All participants provided informed consent. The sports clubs were randomly drawn from the population of soccer and swimming clubs in the urbanized area of two Dutch cities, Amsterdam and Utrecht. All selected clubs agreed to participate. The average size of the young persons’ sections of the clubs was 224 youth members. The participants, 153 male and 107 female athletes (N = 260), were recruited from 25 teams, and ranged from 12 to 18 years of age (M = 14.8, SD = 1.5). The response percentage was high, that is, 90%. Each participant received a 5 Euro CD-token.

Socioeconomic status was a combination of the educa- tional and occupational background of both parents (Van Westerlaak et al., 1990) and was computed on the basis of sample-specific factor loadings and standard deviations. Mean scores correspond to socioeconomic strata in the fol- lowing way: 3 to 9, lower class; 9 to 12, middle class; and 12 to 16, upper class (Bernstein and Brandis, 1970). The internal consistency reliability of the scale for socioe- conomic status was good, α = .82. The mean score was 8.4 (SD = 2.7), which indicated that the sample could be

considered as lower-middle class. The adolescents’ level of formal education was middle to high, and correlated sig- nificantly with the socioeconomic status of their parents, r(235) = .40, p < .01. The mean family size was 2.8 children. The percentage of single parent families was 18.4%, and the percentage of divorced parents was 15.1%. The sample consisted of adolescents with a Dutch (n = 208), Moroccan (n = 20), Surinam (n = 17) and Turkish (n = 15) ethnic back- ground. At the time of the data collection the adolescents had participated in competitive sports during 8.2 years (SD = 3.1) on average.

Measures

The participants completed questionnaires on social desir- ability, anti- and prosocial behavior, sociomoral atmosphere of the sporting environment, sociomoral reasoning about sport dilemmas, and coach-athlete relationship quality. For the purpose of interpretation, all scores were keyed to the names of the scales. For instance, a high score on the scale for antisocial behavior is indicative of a high level of reported antisocial behavior.

Social desirability

The social desirability scale contained 15 items describing socially desirable attributes (Crowne and Marlowe, 1960; Rutten, 2001). Examples are: “I always practice what I preach,” and “I never boast.” Adolescents indicated whether each statement was true or false for them personally. The scale for social desirability proved to be internally consis- tent, α = .83.

Antisocial behavior

Antisocial behavior was assessed with the Anti Social Behavior Inventory (ASBI), which has been used with ado- lescents from the general population and several (sub)clinical samples (Tavecchio et al., 1999; Wouters and Spiering, 1990). The items are concerned with petty crime, vandalism, violence, and rebellious behavior. Corresponding examples of questions are: “I have stolen something that was worth less than 5 euro,” “I have purposely destroyed or damaged things from others,” “I have used a weapon in a fight,” and “I have joined a group in messing around.” The internal con- sistency reliability (15 items) was α = .91. The correlation with social desirability was non-significant.

Prosocial behavior

In order to assess prosocial behavior, we adapted the Proso- cial Behavior Questionnaire (PBQ), and used it as a self- report measure of prosocial behavior in adolescence. The PBQ was designed by Weir and Duveen (1981) in order

Springer

258 J Youth Adolescence (2007) 36:255–264

to be used as a teacher or parent report measure of chil- dren’s prosocial behavior in middle childhood. The PBQ contains 20 items with a 4-point Likert-type scale, ranging from “rarely applies” to “certainly applies.” The items rep- resent positive social behaviors such as helping, sharing and supporting others. Some examples are: “I spontaneously pick up things, which another person has dropped,” and “I take the opportunity to praise the work of those who are less able.”

Stams et al. (2005) used the adapted PBQ in a sample of juvenile delinquents (n = 75) and adolescents from low socioeconomic backgrounds and different cultural minority groups. The PBQ proved to be internally consistent (α = .71). Evidence for concurrent validity of the PBQ was found in significant and positive associations with empathy and victim-based moral orientation, and negative associ- ations with norm-trespassing, delinquent and aggressive behavior. Divergent validity was demonstrated by low to moderate correlations with verbal intelligence, r(652) = .10, p < .05, and social desirability, r(652) = .32, p < .001. In the present study, we found an internal consistency reliability of α = .89. The association with social desirability was r(258) = .22, p < .001.

Sociomoral atmosphere of the sporting environment

The scale for sociomoral atmosphere is an adaptation of the 24-item School Culture Scale (Higgins, 1995, 1997), which has been translated in Dutch by Veugelers and De Kat (1998), who used it in a middle and high school population, showing internal consistency and factorial validity. We adapted the translated version for use in the context of organized youth sport by making the items sport-specific (e.g. athletes re- port on their team-members in stead of their class-mates), removing items that have no meaning in the context of orga- nized youth sport. Subsequently, we conducted a principal components analysis, which showed a one-dimensional so- lution. The remaining 17-items were used as a self report measure that purports to assess the sociomoral climate of a sports club in terms of normative expectations (“There is very little physical fighting”); athlete-coach/sports club re- lationships (“Athletes and coaches trust each other”); and athlete relationships (“Athletes generally treat each other with respect and fairness”). Athletes indicated on 5-point Likert-type scales the degree to which statements regarding the sociomoral climate of their sporting environment were true or untrue. The internal consistency reliability of the scale for sociomoral atmosphere was α = .86. The correlation with social desirability was non-significant.

Sociomoral reasoning about sport dilemmas

The Practical Sociomoral Reflection Objective Measure— Sport (PSROM-Sport) was developed to assess practical so-

ciomoral reasoning in the context of organized youth sport, and was derived from the Sociomoral Reflection Objective Measure—Short Form, the SROM-SF (Basinger and Gibbs, 1987), which is a multiple choice questionnaire contain- ing 2 moral dilemmas and 12 question arrays focusing on sociomoral norms. Each question includes a response op- tion representative of Kohlberg’s moral stages 1 through 4. The first two stages, indicative of unilateral (concrete con- sequences) and instrumental (pragmatic deals or exchanges) reasoning respectively, constitute the immature level. The third and fourth stage, mutual-prosocial and systemic rea- soning respectively, constitute the mature level (Gibbs et al., 1992).

The PSROM-Sport assesses the level of sociomoral rea- soning in a similar way, using 12 question arrays about situations in the context of organized youth sport (e.g., “You decide to help the best player in the team to get fit after an injury, so that he might be ready in time for the most im- portant match of the year”), tapping the type of sociomoral norms the person uses (e.g., “Because this player might help you too” (stage 2), “Without this player you might loose the important match” (stage 1), “It shows that you feel respon- sible for your team (stage 4), “If you don’t, you don’t act as a real friend” (stage 3). The internal consistency reliabil- ity of the PSROM-Sport was sufficient, that is, α = .68. The correlation with social desirability was non-significant.

Some evidence for construct validity was found in a sub-sample of 40 athletes by comparing moral scores on the PSROM-sport with scores on the Sociomoral Reflection Measure—Short Form (SRM-SF), which is a semi-structured interview assessing sociomoral reason- ing competence in general (Gibbs et al., 1992), so not limited to sports related dilemma’s, and a self-constructed semi-structured interview assessing fair play orientation (Loland and McNamee, 2000; Tamboer and Steenbergen, 2000). Both interviews were reliably scored, with intraclass correlations for intercoder agreement above .80, and were internally consistent (α > .80). We found significant correlations between sociomoral reasoning in the context of organized youth sport (PSROM-sport), and both sociomoral reasoning competence, r(38) = .27, p < .05, and fair play orientation, r(38) = .53, p < .001.

Coach-athlete relationship quality

The Quality of the Relationship scale is an adaptation of the Barrett-Lennard Relationship Inventory (Barrett-Lennard, 1962), and was construed to assess the quality of the relationship between the athlete and his or her coach in terms of empathic understanding, positive regard, congru- ence, and willingness to be known. These adjectives are considered to be features of a rewarding coach-athlete rela- tionship (Poczwardowski et al., 2002; Van den Auwele and

Springer

J Youth Adolescence (2007) 36:255–264 259

Rzewnicki, 2000). By means of 6-point Likert–type scales, ranging from total disagreement to total agreement, adoles- cents responded to 12 statements concerning their present relationship with the coach. Athletes responded to items such as “I appreciate this coach” and “I nearly always know exactly what this coach means.” The internal consistency reliability of the Quality of Relationship scale was α = .87. The correlation with social desirability was non-significant.

Results

Descriptive analyses

Table 1 presents the correlations between age, level of education, cultural background (Caucasian white or cultural minority), socioeconomic status, extent of sport participa- tion (the standardized summation of the number of hours and days per week spent in sporting activity), type of sport (soccer or swimming), type of team (male or female ath- letes), social desirability, sociomoral reasoning, sociomoral atmosphere, coach-athlete relationship quality, and anti- and prosocial behavior. Only effects at p < .001 were considered significant in order to adjust for multiple comparisons.

Age correlated positively with sociomoral reason- ing, r(258) = .25, p < .001. The extent of sport partic- ipation was less in swimmers than in soccer players, r(258) = − .20, p < .001, and less in female teams than in male teams, r(258) = − .27, p < .001. Athletes from cul- tural minority groups were underrepresented in compet- itive swimming, r(258) = − .30, p < .001, and in female teams, r(258) = − .24, p < .001. Socioeconomic status was positively associated with level of education, r(258) = .37, p < .001, and negatively associated with social desirability, r(258) = − .28, p < .001. Also, level of education was nega- tively related to social desirability, r(258) = − .28, p < .001. Social desirability proved to be associated with prosocial behavior, r(258) = .22, p < .001.

Sociomoral atmosphere was evaluated as relatively pos- itive by competitive swimmers, r(258) = .29, p < .001, and female athletes, r(258) = .30, p < .001. Higher coach-athlete relationship quality was related to more mature levels of so- ciomoral reasoning, r(258) = .21, p < .001, and positive per- ceptions of sociomoral atmosphere, r(258) = .45, p < .001.

A positive sociomoral atmosphere, r(258) = − .39, p < .001, and higher coach-athlete relationship quality, r(258) = .38, p < .001, were negatively associated with antisocial behavior, and positively associated with prosocial behavior, r(258) = .33, p < .001, and r(258) = .26, p < .001, respectively. Moral reasoning was only associated with prosocial behavior, r(258) = .22, p < .001. The incidence of antisocial behavior was relatively low in competitive swimmers, r(258) = − .20, p < .001. Finally, female ath- letes reported more prosocial behavior than male athletes, r(258) = .26, p < .001.

T ab

le 1

M ea

ns ,s

ta nd

ar d

de vi

at io

ns ,a

nd co

rr el

at io

ns be

tw ee

n ba

ck gr

ou nd

va ri

ab le

s, so

ci al

de si

ra bi

li ty

,e xp

la na

to ry

va ri

ab le

s, an

d an

ti -a

nd pr

os oc

ia l

be ha

vi or

V ar

ia bl

es M

S D

1 2

3 4

5 6

7 8

9 10

11 12

13

1. A

ge 14

.7 8

1. 54

1. 00

2. L

ev el

of ed

uc at

io n

3. 86

2. 07

.0 1

1. 00

3. C

ul tu

ra l

ba ck

gr ou

nd 1.

25 .4

3 −.

05 −.

10 1.

00 4.

S E

S 8.

41 2.

60 .1

2 .3

7∗ −.

15 1.

00 5.

E xt

en t

of pa

rt ic

ip at

io n

.0 0

.8 7

.0 9

−. 05

−. 01

−. 01

1. 00

6. T

yp e

of sp

or t

.2 8

.4 5

.0 1

.0 1

−. 30

∗ −.

11 −.

20 ∗

1. 00

7. T

yp e

of te

am .4

1 .4

9 −.

04 .1

1 −.

24 ∗

.0 9

−. 27

∗ .1

2 1.

00 8.

S oc

ia l

de si

ra bi

li ty

1. 33

.2 5

−. 17

−. 28

∗ .1

7 −.

22 ∗

.0 4

−. 11

−. 11

1. 00

9. M

or al

re as

on in

g 2.

75 .4

4 .2

5∗ .1

2 −.

13 .0

9 .0

6 .0

4 .1

7 −.

08 1.

00 10

.M or

al at

m os

ph er

e 3.

73 .5

3 −.

15 .1

1 −.

16 .0

7 −.

09 .2

9∗ .3

0∗ .0

5 .0

7 1.

00 11

.R el

at io

ns hi

p qu

al it

y 4.

71 .7

5 −.

03 .0

8 −.

10 .0

7 .0

0 .0

3 .1

1 .0

2 .2

1∗ .4

5∗ 1.

00 12

.A nt

is oc

ia l

be ha

vi or

1 .3

2 .3

2 .1

3 −.

15 .1

8 −.

03 .0

8 −.

20 ∗

−. 16

−. 07

−. 11

−. 39

∗ −.

38 ∗

1. 00

13 .P

ro so

ci al

be ha

vi or

2. 75

.4 4

−. 01

−. 01

−. 03

.0 0

−. 01

.0 1

.2 6∗

.2 2∗

.2 2∗

.3 3∗

.2 6∗

−. 19

1. 00

N o

te . N

= 26

0. ∗ p

< .0

01 .

Springer

260 J Youth Adolescence (2007) 36:255–264

Multilevel analysis

Multilevel analysis allows the simultaneous examination of how individual and group level variables are related to indi- vidual level outcomes, accounting for the non-independence of observations within groups (Goldstein, 1995). Using mul- tilevel regression analysis, we separately tested a model for antisocial and prosocial behavior, examining whether so- ciomoral atmosphere, sociomoral reasoning, coach-athlete relationship quality (explanatory variables) predicted antiso- cial and prosocial behavior, controlling for social desirability and background variables, such as sex, age and type of sport.

The explanatory variables could be considered as quali- ties of individual athletes and as features of the social con- text. Therefore, all explanatory variables were analyzed both at the individual and contextual level. In order to separate individual effects from contextual effects, we aggregated characteristics of individual athletes within teams, and sub- sequently subtracted the group mean from the corresponding individual scores. By this method of group mean centering individual level variables were derived, having analogues at the contextual level. The individual level variables reflect individual variation in perceived sociomoral atmosphere, so- ciomoral reasoning, and perceived coach-athlete relationship quality. The contextual level variables reflect environmen- tal variation or between-team differences in sociomoral at- mosphere, sociomoral reasoning, and coach-athlete relation- ship quality. Apart from the derived contextual variables, we distinguished integral contextual variables, such as type of sport (soccer versus competitive swimming) and type of team (male versus female athletes), which are not summaries of individual characteristics and have no individual level equivalents.

We tested multilevel regression models for antisocial and prosocial behavior in three consecutive steps. Model fit was tested by the difference in deviance between each step, which has a chi-square distribution, and can be used to test whether the more elaborate model fits significantly better than the simpler model. The variables with non-significant coeffi- cients were removed from the model after each step. The resulting models were used as a reference for further com- parison. We chose to present only the best fitting multilevel regression models for antisocial and prosocial behavior.

In the first step, a null-model or intercept-only model that contains an outcome variable and no explanatory vari- ables was fitted to the data as a baseline. In the second step, individual level explanatory variables were entered in the model, such as perceived coach-athlete relationship quality, controlling for social desirability and all background variables. This step deals with perceptions, cognitions and experiences of individual athletes. In the third step, contex- tual explanatory variables were entered in the model, such as mean coach-athlete relationship quality, and both integral

variables, namely, type of sport (soccer versus competitive swimming), and type of team (male versus female). As such, we took into account that athletes were always part of spe- cific teams. In this way the perspective was shifted from the individual athlete to the immediate context of athletic action, namely, the team or the coach. Mean sociomoral atmosphere is a shared perspective, and could be considered as a more objective measure of the sociomoral atmosphere. Mean so- ciomoral reasoning reflects the team’s level of sociomoral reasoning. Mean coach-athlete relationship quality reflects the degree to which coaches are able to maintain good rela- tionships with their athletes.

Multilevel regression model for antisocial behavior

From the null-model (see Table 2) it can be derived that 92% of the variance in antisocial behavior could be attributed to differences among individual athletes within teams (in- dividual level), and that 8% of the variance in antisocial behavior could be attributed to differences between teams or coaches (contextual level). The best fitting multilevel regres- sion model—X2(5, N = 260) = 63.93, p < .001–accounted for 23% of the variance in antisocial behavior among athletes. Most of the variance accounted for in antisocial behavior was distributed at the individual level, namely, 15%. The explained variance at the contextual level was 8%. Higher perceived and mean coach-athlete relationship quality were related to lower levels of antisocial be- havior, b = − .20, t(259) = 3.36, p < .001, and b = − .22, t(259) = 4.13, p < .001, which indicated that more positive coach-athlete relationships predicted lower levels of anti- social behavior both at the individual and contextual level. Athletes who perceived their sporting environment to be rel- atively positive reported lower levels of antisocial behav- ior, b = − .21, t(259) = 3.49, p < .001. Type of sport was negatively associated with antisocial behavior, b = − .18, t(259) = 3.37, p < .001, indicating that swimmers scored rel- atively low on antisocial behavior in comparison with soccer players. Athletes higher in formal education reported less antisocial behavior, b = − .11, t(259) = 2.13, p < .05. We found no significant individual-level or cross-level interac- tion effects.

Multilevel regression model for prosocial behavior

The null-model in Table 3 shows that 93% of the variance in prosocial behavior could be attributed to differences among individual athletes within teams, and that 7% of the variance in prosocial behavior could be attributed to differences between teams or coaches. The best fitting multilevel regres- sion model—X2(5, N = 260) = 62.28, p < .001—accounted for 23% of the variance in prosocial behavior among athletes. Most of the variance in prosocial behavior that

Springer

J Youth Adolescence (2007) 36:255–264 261

Table 2 Multilevel regression model for antisocial behavior

Explanatory model Null-model b t

Individual level Perceived coach-athlete relationship quality − .20 3.36∗∗∗ Perceived sociomoral atmosphere − .21 3.49∗∗∗ Level of education − .11 2.13∗

Contextual level Mean coach-athlete relationship quality − .22 4.13∗∗∗ Type of sport (soccer vs. swimming) − .18 3.37∗∗∗

Variance components Contextual level .008 (8%) .000 Individual level .091 (92%) .076

Explained variancea

Contextual level 8% Individual level 15%

Deviance 131.023 67.096 X2 63.93∗∗∗

Note. N = 260 athletes, N = 25 teams. aTotal amount of explained variance is 23%. ∗p < .05; ∗∗∗p < .001.

was explained by the independent variables was distributed at the individual level, namely, 16%, while 7% of the explained variance was distributed at the contextual level. Athletes displaying a stronger tendency to respond in a socially desirable manner reported more prosocial behavior, b = .24, t(259) = 4.19, p < .001. Higher levels of sociomoral reasoning were associated with more prosocial behavior both at the contextual level, b = .15, t(259) = 2.23, p < .05, and individual level, b = .17, t(259) = 3.03, p < .01. Also, a relatively positive perception of the sociomoral atmosphere

was associated with more prosocial behavior, b = .26, t(259) = 4.56, p < .001. Finally, female teams reported higher levels of prosocial behavior, b = .20, t(259) = 3.03, p < .01. No significant interaction effects were found.

Discussion

In this study, we focused on educationally relevant factors in organized youth sport that were hypothesized to contribute to

Table 3 Multilevel regression model for prosocial behavior

Explanatory model Null-model b t

Individual level Sociomoral reasoning .17 3.03∗∗

Perceived sociomoral atmosphere .26 4.56∗∗∗

Social desirability .24 4.19∗∗∗

Contextual level Mean sociomoral reasoning .15 2.23∗

Type of team (male vs. female) .20 3.03∗∗

Variance components Contextual level .013 (7%) .000 Individual level .179 (93%) .149

Explained variancea

Contextual level 7% Individual level 16%

Deviance 304.525 242.245 X2 62.28∗∗∗

Note. N = 260 athletes, N = 25 teams. aTotal amount of explained variance is 23%. ∗p < .05; ∗∗p < .01; ∗∗∗p < .001.

Springer

262 J Youth Adolescence (2007) 36:255–264

antisocial and prosocial behavior of adolescent athletes. The sample consisted of 260 male and female soccer players and competitive swimmers. Using multilevel regression analysis, we found that 8% of the variance in antisocial behavior and 7% of the variance in prosocial behavior could be attributed to the sporting environment, more specifically, the team and its coach. We examined whether sociomoral atmosphere of the sporting environment, sociomoral reasoning about sport dilemmas, and coach-athlete relationship quality could explain these environmental effects, independent of type of sport and important characteristics of the athletes, including age, sex, level of education, current extent and years of sport participation, cultural background, and socioeconomic status.

Sociomoral reasoning was significant as a contextual fac- tor explaining the influence of team-membership on proso- cial behavior, and coach-athlete relationship quality was significant as a contextual factor explaining the influence of team-membership on antisocial behavior. These findings are in line with research emphasizing the important role played by the peer group (Carlo et al., 1999), and studies showing the importance of relationship quality with par- ents or other important adults (Scholte et al., 2001; Stattin and Kerr, 2000). At the individual level, more explanatory factors were found. Athletes who experienced a favorable sociomoral atmosphere of the sporting environment and a positive relationship with their coach reported less antisocial behavior. More prosocial behavior was predicted by posi- tive perceptions of the sociomoral atmosphere, and mature sociomoral reasoning.

Multilevel research in the context of the school showed that 19% of the variance in scholastic achievement among students was attributed to characteristics of the school en- vironment (Scheerens and Bosker, 1997). In the present study, 8% of the variance in antisocial behavior and 7% of the variance in prosocial behavior could be attributed to characteristics of the sporting environment. These percent- ages may not appear to be large. However, we must take into account that our study is concerned with a cross-context prediction of antisocial and prosocial behavior. We probably would have found larger percentages of contextual variation if we had focused on behaviors within the direct context of athletic action, that is, antisocial and prosocial behavior on the playing field, but this was not the purpose of our study. The contextual effects that we found seem relatively small, partly because individual effects tend to be systematically overestimated at the expense of contextual effects. Notably, in multilevel analysis all measurement error is exclusively distributed at the individual level.

In the present study, contextual effects provide the strongest evidence for the influence of coach-athlete rela- tionship quality and sociomoral reasoning on antisocial and prosocial behavior, as these effects are independent from the

athlete’s contribution to the team and the relationship with the coach. It is possible that contextual effects can be explained by mechanisms of self-selection. One such mechanism is the process that athletes may choose to become members of a specific team because they share the same personal quali- ties as other athletes. Another example of self-selection is the process by which athletes with similar characteristics are allocated to specific teams or coaches.

Selective allocation may largely explain why female teams scored higher on prosocial behavior than male teams, because the contextual effect for type of team could be a direct result of a normative gender effect on measures of prosocial behavior in adolescents (Carlo and Randall, 2002; Eagly and Crowly, 1986). More prosocial behavior in female athletes is in line with research showing that female athletes reported higher levels of moral functioning, lower approval of unsportsmanlike behaviors, and were less likely to judge injurious acts as legitimate than male athletes (Kavussanu and Roberts, 2001).

There are two arguments against self-selection. Firstly, we found no effect for type of sport on prosocial behavior. Secondly, the contextual effects for coach-athlete relation- ship quality and sociomoral reasoning remained significant after controlling for type of sport and other variables that may be connected with processes of self-selection, including for- mal education, age, sex, socioeconomic status, and cultural background. Although we should be extremely careful about making any causal inferences, these findings render the argu- ment for self-selection less compelling as an explanation for the contextual effects found in the domains of coach-athlete relationship quality and sociomoral reasoning.

There are some limitations to the current study. The most important limitation is the cross-sectional nature of the de- sign, which sets limits to the causal interpretation of our findings. This limitation was partly overcome by using multi- level regression analyses. However, as we already discussed, selection effects cannot be ruled out completely, because not all potential confounding variables can ever be statistically controlled. Therefore, we should caution against the causal interpretation of our results, especially if individual effects are involved. Although the contextual effects suggest a causal link between sport participation and antisocial and prosocial behavior, the contextual effects of coach-athlete relationship quality and sociomoral reasoning should be corroborated in a prospective study using an experimental design. A second limitation is that our findings are based on measures of self- report. Hence, we cannot show to what extent self-report of antisocial and prosocial behavior was contaminated by in- dividual perceptions of athletes. However, there is evidence for the reliability and validity of self-report measures assess- ing antisocial (Junger-Tas and Haen Marshall, 1999) and prosocial behavior (Carlo and Randall, 2002). Because self- report of antisocial and prosocial behavior may be sensitive to

Springer

J Youth Adolescence (2007) 36:255–264 263

responding in a socially desirable manner, as was the case for prosocial behavior, we controlled for social desirability in the multilevel regression analyses.

This study is the first to examine how individual and contextual characteristics of the sporting environment might foster antisocial and prosocial behavior in adolescent athletes. Results suggest that coaches who maintain good relationships with their athletes reduce antisocial behavior, and that exposure to relatively high levels of sociomoral rea- soning within the immediate context of sporting activities promotes prosocial behavior. We conclude that the context of organized youth sport contains moral and relational fac- tors that are relevant to understanding sports’ contribution to adaptive behavioral development of young sportsmen and sportswomen.

References

Anderson CA, Morrow M (1995) Competitive aggression without in- teraction: effects of competitive versus cooperative instructions on aggressive behavior in video games. Pers Soc Psychol Bull 21:1020–1030

Arnold PJ (1994) Sport and moral education. J Moral Educ 23:75–89 Arnold PJ (2001) Sport, moral development, and the role of the teacher:

Implications for research and moral education. Quest 53:135– 150

Barrett-Lennard GT (1962) Dimensions of therapist response as causal factors in therapeutic change. Psychol Monogr 76(43):562

Basinger KS, Gibbs JC (1987) Validation of the sociomoral reflection objective measure-short form. Psychol Rep 61:139–146

Begg DJ, Langley JD, Moffitt T, Marshall SW (1996) Sport and delin- quency: an examination of the deterrence hypothesis in a longitu- dinal study. Br J Sports Med 30:335–341

Bernstein B, Brandis W (1970) Social class differences in communi- cation and control. In: Brandis W (ed) Social class, language, and communication. Routledge, London, pp 93–124

Blasi A (1980) Bridging moral cognition and moral action: a critical review of the literature. Psychol Bull 88:1–45

Bredemeier BJL (1994) Children’s moral reasoning and their assertive, aggressive, and submissive tendencies in sport and daily life. J Sport Exerc Psychol 16:1–14

Carlo G, Fabes RA, Laible D, Kupanoff K (1999) Early adolescence and prosocial/moral behavior II: the role of social and contextual influences. J Early Adolesc 19:133–147

Carlo G, Randall BA (2002) The development of a measure of prosocial behaviors for late adolescents. J Youth Adolesc 31:31–44

CBS (1999) Sportbeoefening vanaf 1991 (Sports participation from 1991) CBS, Voorburg

Coakley JJ (1984) Mead’s theory on the development of the self: im- plications for organized youth sport programs. Paper presented at the Olympic Scientific Congress, Eugene, OR

Coalter F (1989) Sport and antisocial behaviour: a literature review. Scottish Sports Council, Edinburgh

Colby A, Kohlberg L, Abrahami A, Gibbs JC, Higgins A, Kauffman K, Lieberman M, Nisan M, Reimer J, Schrader D, Snarey J, Tappan MB (1987). The measurement of moral judgment. Volume I, Theo- retical foundations and research validation. Cambridge University Press, Cambridge, NY

Conroy DE, Silva JM, Newcomer RR, Walker BW, Johnson MS (2001) Personal and participatory socializers of the perceived legitimacy of aggressive behavior in sport. Aggressive Behav 27:405–418

Cotterell J (1996) Social networks and influences in adolescence. Rout- ledge, London and New York

Crowne DP, Marlowe D (1960) A new scale of social desirability inde- pendent of psychopathology. J Consult Psychol 24:349–354

Darling N, Hamilton SF, Toyokawa T, Matsuda S (2002) Naturally occurring mentoring in Japan and the United States: social roles and correlates. Am J Community Psychol 30:245–270

DuBois DL, Holloway BE, Valentine JC, Cooper H (2002) Effective- ness of mentoring programs for youth. A meta-analytic review. Am J Community Psychol 30:157–197

Duncan SC, Duncan TE, Strycker LA, Chaumeton NR (2002) Relations between youth antisocial and prosocial activities. J Behav Med 25:425–438

Eagly AH, Crowly M (1986) Gender and helping behavior: a meta- analytic review of the social psychological literature. Psychol Bull 100:283–308

Emler N, Reicher S (1995) Adolescence and delinquency: the collective management of reputation. Blackwell Publishers, Cambridge, MA

Gibbs JC, Basinger KS, Fuller D (1992) Measuring the development of sociomoral reflection. Lawrence Erlbaum, Hillsdale, NY

Goldstein H (1995) Multilevel statistical models, 2nd edn. Edward Arnold, London

Guivernau M, Duda JL (2002) Moral atmosphere and athletic aggres- sive tendencies in young soccer players. J Moral Educ 31:67– 85

Higgins A (1995) Dimensions of the school culture scale: measuring attitudes, norms and values in educational settings. Fordham Uni- versity, Fordham

Higgins A (1997) Research using the school culture scale. Int J Educ Res 27:558–564

Junger-Tas J, Haen Marshall I (1999) The self-report methodology in crime research. In: Tonry M (ed) Crime and justice. A re- view of research, vol. 25. University of Chicago Press, Chicago, pp 291–367

Kavussanu M, Ntoumanis N (2003) Participation in sport and moral functioning: does ego-orientation mediate their relationship? J Sport Exerc Psychol 25:501–518

Kavussanu M, Roberts GC (2001) Moral functioning in sport: an achievement goal perspective. J Sport Exerc Psychol 1:37– 54

Kleiber DA, Roberts GC (1981) The effects of sport experience in the development of social character: a preliminary investigation. J Sport Exerc Psychol 3:114–122

Kohn A (1986) No contest: the case against competition. Houghton Mifflin, Boston

Kreft I, De Leeuw J (1998) Introducing multilevel modeling. Sage, London, Thousand Oaks, New Delhi

Loland S, McNamee M (2000) Fair play and the ethos of sports: an eclectic philosophical framework. J Philos Sport 27:63– 80

Mahoney JL (2000) School extracurricular activity participation as a moderator in the development of antisocial patterns. Child Dev 71:502–516

Mahoney JL, Stattin H (2000) Leisure activities and adolescent antiso- cial behavior: the role of structure and social context. J Adolesc 23:113–127

Matsuba MK, Walker LJ (1998) Moral reasoning in the context of ego functioning. Merill-Palmer Q 44:464–483

Mutrie N, Parfitt G (1998) Physical activity and its link with mental, social and moral health in young people. In: Biddle BS, Sallis J, Cavill N (eds) Young and active? Young people and health- enhancing physical activity—Evidence and implications. Health Education Authority, London, pp 49–68

Nelson JR, Smith JJ, Dodd J (1990) The moral reasoning of juvenile delinquents: a meta-analysis. J Abnorm Child Psychol 18:231– 239

Springer

264 J Youth Adolescence (2007) 36:255–264

Ommundsen Y, Roberts GC, Lemyre PN, Treasure D (2003) Perceived motivational climate in male youth soccer: relations to social- moral functioning, sportspersonship and team norm perceptions. Psychol Sport and Exerc 4:397–413

Poczwardowski A, Barott JE, Henschen KP (2002) The athlete and coach: their relationship and its meaning. Results of an interpretive study. Int J Sport Psychol 33:116–140

Power FC, Higgins A, Kohlberg L (1989) Lawrence Kohlberg’s ap- proach to moral education. Columbia University Press, New York

Rhodes JE (2002) Stand by me: the risks and rewards of mentoring today’s youth. Harvard University Press, Cambridge, MA

Robins D (1990) Sport as prevention. The role of sport in crime pre- vention programmes aimed at young people. Centre for Crimino- logical Research, Oxford University, Oxford

Rutten EA (2001) De morele oriëntatie lijst: Een valideringstudie (The moral orientation list: a validation study). Leiden University, Leiden

Scheerens J, Bosker RJ (1997) The foundations of educational effec- tiveness. Pergamon, UK

Scholte RHJ, Van Lieshout CFM, Van Aken MAG (2001) Perceived relational support in adolescence: dimensions, configurations, and adolescent adjustment. J Res Adolesc 11:71–94

Shields DLL, Bredemeier BJL (1995) Character development and physical activity. Human Kinetics, Champaing, IL

Snijders TAB, Bosker RJ (1999) Multilevel analysis: an introduction to basic and advanced multilevel modeling. Sage, London, Thousand Oaks, New Delhi

Stams GJJM, Dekovic M, Brugman D, Rutten EA, Van den Wittenboer GLH, Tavecchio, LWC, Hendriks J, Van Schijndel M (2006). The relation of punishment- and victim-based moral orientation to prosocial, externalizing, and norm trespassing behavior in delin- quent and non-delinquent adolescents

Stattin H, Kerr M (2000) Parental monitoring: a reinterpretation. Child Dev 71:1072–1085

Stephens DE (2000) Predictors of likelihood to aggress in youth soccer: an examination of coed and all-girls teams. J Sport Behav 23:311– 325

Stephens DE, Bredemeier BJL (1996) Moral atmosphere and judgments about aggression in girls’ soccer: relationships among moral and motivational variables. J Sport Exerc Psychol 18:158–173

Tamboer J, Steenbergen J (2000) Sportfilosofie (Philosophy of sport). Damon, Leende

Tavecchio LWC, Stams GJJM, Brugman D, Thomeer-Bouwens MAE (1999) Moral judgement development and delinquency in home- less youth. J Moral Educ 28:63–79

Vallerand RJ, Deshaies P, Cuerrier JP (1997) On the effects of the social context on behavioral intentions of sportsmanship. Int J Sport Psychol 28:26–140

Van Bottenburg M, Schuyt K (1996) De maatschappelijke beteke- nis van sport (The societal meaning of sport). NOC ∗ NSF, Arnhem

Van den Auwele Y, Rzewnicki R (2000) Putting relationship issues in sport in perspective. Int J Sport Psychol 31:573–577

Van Westerlaak JM, Kropman JA, Collaris JWM (1990) Beroepenklap- per (Standard educational and occupational classification guide). ITS, Nijmegen

Vazsonyi AT, Pickering LE, Belliston LM, Hessing D, Junger M (2002) Routine activities and deviant behaviors: American, Dutch, Hungarian, and Swiss Youth. J Quant Criminol 18:397– 422

Veugelers W, De Kat E (1998) Opvoeden in het voortgezet onderwijs: Leerlingen, ouders en docenten over de pedagogische opdracht en de afstemming tussen gezin en school (Education in middle and high school: students, parents and teachers about their educational tasks in the attunement between family and school.). Van Gorcum, Assen

Weir K, Duveen G (1981) Further development and validation of the prosocial behaviour questionnaire for use by teachers. J Child Psychol Psychiatry 22:357–374

Weiss MR, Smith AL (2002) Friendship quality in youth sport: rela- tionship to age, gender, and motivation variables. J Sport Exerc Psychol 24:420–437

Wills TA, Cleary SD (1996) How are social support effects mediated? A test with parental support and adolescent substance use. J Pers Soc Psychol 71:973–952

Wouters L, Spiering W (1990) Vragenlijst ongewenst gedrag (Antiso- cial behavior questionnaire). GG & GD Amsterdam, Amsterdam

Zimmerman MA, Bingenheimer JB, Notaro PC (2002) Natural men- tors and adolescent resilience: a study with urban youth. Am J Community Psychol 30:221–243

Springer

COPYRIGHT INFORMATION

TITLE: The Contribution of Organized Youth Sport to Antisocial and Prosocial B

SOURCE: Journal of Youth and Adolescence 36 no3 Ap 2007 PAGE(S): 255-64

The magazine publisher is the copyright holder of this article and it is reproduced with permission. Further reproduction of this article in violation of the copyright is prohibited. To contact the publisher: http://www.springerlink.com/content/1573-6601/