behavior in juveniles
A meta-analytic review
Master’s thesis Forensische Orthopedagogiek, Graduate School of Child Development and Education
University of Amsterdam
Student: J. Gubbels Student number: 10625046 First supervisor: T. van der Stouwe MSc Second supervisor: Prof. dr. G.J.J.M. Stams Amsterdam, July 2015
2
Abstract
Martial arts are very popular among juveniles all over the world, but the effects of martial arts on externalizing behavior in juveniles remains unclear. The current multilevel meta-analysis of 9 studies, including 76 effect sizes and N = 1169 juveniles, was conducted to examine the relation between martial arts participation and externalizing behavior in
juveniles, and to examine factors with a possible moderating effect. Results showed that there was no overall effect of martial arts participation on externalizing behavior in juveniles. However, in comparison to non-athletes and team sport athletes, less externalizing behavior was found in the martial artists. Furthermore, judokas specifically showed more externalizing behavior, whereas juvenile participants in other martial arts showed a similar level or less externalizing behavior. Implications for future policy and research are discussed.
3 Table of Contents Abstract 2 Introduction 4 Method 8 Selection of Studies 8 Inclusion Criteria 8
Coding the Studies 9
Data Analysis 9 Publication Bias 10 Results 11 Overall Effects 11 Moderator Effects 14 Multivariate Analysis 15 Discussion 16 Limitations 29
Strengths and Implications for the Future 20
Conclusion 21
References 22
4
Martial arts participation and externalizing behavior in juveniles: a meta-analytic review
In the past century, sports have earned an important function in society because of their supposed positive influence on the development of juveniles. Sports psychologists have argued that life skills can be taught in combination with athletic skills in sport contexts, and this belief has contributed to the implementation of sports in several life skill programs (e.g., Danish, Nellen, & Owens, 1996; Danish, Petitpas, & Hale, 1995). Through these programs better goal setting, improved problem solving skills and positive thinking in children have been achieved (Papacharisis, Goudas, Danish, & Theodorakis, 2005).
In addition to the positive effects of sport- based programs, research has shown that practicing sports in general is positively related to school involvement (Marsh & Kleitman, 2003), self-esteem (Crews, Lochbaum, & Landers, 2004; Daniels & Leaper, 2006) and social competence (Hedstrom & Gould, 2004; Ullrich-French, McDonough, & Smith, 2012).
Furthermore, the catharsis theory claims that sport participation will vent off negative energy, like aggression and other externalizing behaviors (Breuer & Freud, 1974). However, some claim that sports could also have negative effects on the development of juveniles. Sports have for instance been associated with higher stress levels, heightened peer pressure and inappropriate adult behavior towards juveniles (i.e., controlling and manipulative behavior by coaches or parents) (Hansen, Larson, & Dworkin, 2003; Danish, Kleiber, & Hall, 1987; Scanlan, Babkes, & Scanlan, 2005; Smoll & Smith, 1996). Furthermore, some authors claim that sports would increase the level of aggressiveness in juveniles (Bredemeier, Shields, Weiss, & Cooper, 1986; Burton & Marshall, 2005), whereas others argue that sport
participation would have no effect on juvenile delinquency (Spruit, Van Vugt, Van der Put, Van der Stouwe, & Stams, 2015). In sum, previous research is inconsistent with respect to the positive influence of sports on the development of juveniles.
5
distinction between different levels of contact sport (Bredemeier et al., 1986; Conroy, Silva, Newcomer, Walker, & Johnson, 2001; Endresen & Olweus, 2005). In most of these studies, (high) contact sports have been associated with negative behavioral outcomes in comparison to (low- or) non-contact sports. In general, the definition of contact sports is associated with combat sports and/or martial arts, so that the negative outcomes are often attributed to these sports specifically. However, the definition of contact sports in current research has ranged from individual combat sports, like wrestling and judo (see e.g., Bredemeier et al., 1986) and individual power sports, like weightlifting (Endresen & Olweus, 2005), to team sports like basketball, field hockey, and soccer (see e.g., Conroy et al., 2001). Consequently, more research is needed to conclude that contact sports and combat sports specifically indeed lead to negative outcomes.
Martial arts (e.g., karate, judo, taekwondo, boxing) are such contact sports that are very popular among children and adolescents all over the world. In most countries martial arts are even the most popular extracurricular sports practiced by children between 10 and 15 years old (De Knop, Engström, Skirstad, & Weiss, 1996). In the Netherlands, approximately 50% to 56% of those who practice martial arts in a club setting, are juveniles (Breedveld, Kamphuis, & Tiessen-Raaphorst, 2008) and in Finland this proportion even ranges from 64% to 86% (Tammelin, Näyhä, Hills, & Järvlin, 2003). In light of the increased popularity of martial arts, it is noteworthy that some controversy exists about the organization and practice of martial arts. Proponents of martial arts consider them as valuable sports with an important function for society. Through martial arts, juveniles would learn to deal with rules and to have respect towards others (Van Koolwijk, Lagendijk, & Van Egdom, 2014). On the other hand, the practice of martial arts raises ethical concerns because of the possible harmful effects on the personal and social well-being of young participants in terms of aggressive and violent behavior (Endresen & Olweus, 2005).
6
Several researchers have studied the effects of martial arts on behavioral outcomes in juveniles. Martial arts have been associated with higher self-esteem (Finkenberg 1990; Trulson, 1986), emotional stability (Konzay & Bourdreau, 1984), assertiveness (Konzak & Boudreau, 1984) and less anxiety (Kurian, Caterino, & Kulhavy, 1993). Moreover, several studies have examined the influence of martial arts on externalizing behavior, and have found positive effects for martial arts (e.g., Daniels & Thornton, 1990; Daniel & Thornton, 1992; Lamarre & Nosanchuk, 1999; Nosanchuk, 1981; Rutten et al., 2011; Steyn & Roux, 2009). Some studies have associated these positive effects with more experience in martial arts (Lakes & Hoyt, 2004; Skelton, Glynn, & Berta, 1991) and specific target groups, such as middle or elementary school juveniles at risk (Edelman, 1994; Zivin et al., 2001). However, not all studies have found positive effects of martial arts on externalizing behavior. Higher levels of externalizing behavior among martial artists have also been reported (Endersen & Olweus, 2005; Lotfian, Ziaee, Amini, & Mansournia, 2011; Reynes & Lorant, 2001; 2002a; 2002b; 2004; Ziaee, Lotfian, Amini, Mansournia, & Mernari, 2012).
The inconsistency in the findings on the effects of martial arts on externalizing behavior may be due to some contextual factors. First, the experience in martial arts training is a factor in predicting aggressiveness, with more experienced martial artists showing less aggression (see Skelton et al., 1991). Second, the type of the practiced combat sport has been associated with differences in aggression. Several authors have found that those who
practiced modern martial arts training (e.g., boxing) showed more aggressiveness than traditionally trained martial artist (e.g., karate) (Graczyk, Hucinski, Norkowski, Peczak-Graczyk & Rozanowska, 2010; Nosanchuk & MacNeil, 1989, Trulson, 1986).
Furthermore, the existing research on the effects of martial arts on externalizing behavior in juveniles shows some important limitations. First of all, a control group is often lacking (Kellogg, 2003; Skelton et al., 1991) so that levels of externalizing behaviors in
7
martial artists could not be compared to athletes in other sports or in non-athletes. Moreover, studies that have made use of a control group often lack definition of the level and type of sports participation in the control group (e.g., Endresen & Olweus, 2005), or applied martial arts by means of an intervention (e.g., Delva-Tauiliili, 1995; Palermo et al., 2006). As martial arts are often participated for a short amount of time in these studies, the results cannot just be generalized to all martial artists.
With the increasing popularity of martial arts in juveniles, more research on the effects of martial arts in juveniles is important. To paint a bigger picture of the effects of martial arts on externalizing behavior in juveniles and to examine the impact of possible moderators on these effects, a level meta-analysis was conducted. Through a multi-level meta-analysis multiple effects within studies can be included to generate more statistical power and to fully exploit all possible factors that may moderate the relation between martial arts participation and externalizing behavior in the available research. Furthermore, little is known about characteristics that influence the effect of martial arts on externalizing behavior in juveniles, because these characteristics mostly do not vary within studies, but between studies. Therefore, several moderators have been examined in the current meta-analysis. These moderators included study characteristics, such as the impact factor of the journal, design characteristics, including the control group assignment and the way martial arts were implemented (i.e., as an intervention or not), sample characteristics, such as age, gender, the type of martial arts, the amount of experience in martial arts and the intensity of training in martial arts, the control group characteristics, including the type of control group (non- athletes or those practicing other sports) and the type of sports practiced in the control group, and the outcome characteristics, including the construct category (aggressive behavior, aggressive attitude and both/other) .
8
In the current meta-analysis, the following research questions will be addressed: 1) What are the effects of martial arts on externalizing behavior? 2) Which study, design,
sample, control group and outcome characteristics have a moderating effect on the outcomes?
Method Selection of Studies
All reports until 2014 addressing the effects of martial arts on externalizing behavior for juveniles were included. Multiple computerized databases have been searched to identify relevant studies: PiCarta, ProQuest Sociological Abstracts, Google Scholar, EBSCOhost Academic Search Premier, Wiley Online Library, OvidSP, National Academic Research and Collaboration Information System (NARCIS), ScienceDirect and Web of Science. The following keywords were used: ‘combat sport’, ‘fight sport’, ‘martial art’, aikido, bojutsu, boxing, capoeira, dambe, eskrima, fencing, gatka, grappling, hapkido, jiu-jitsu, judo, jujutsu, kalaripayattu, karate, kendo, kenjutsu, kickboxing, kobudo, ‘krav maga’, ‘kung fu’, kyudo, ‘muay thai’, pankration, sambo, sanda, sanshou, savate, self-defence, shootboxing, silat, sumo, taekwondo, ‘thai chi’, ‘vale tudo’, ‘wing chun’, wrestling and wushu. To limit the amount of results, some of these keywords were combined with the keyword ‘sport’. To obtain possible unpublished and non-reported data on study and sample characteristics and outcomes, several authors were contacted.
Inclusion Criteria
To be included in the current meta-analysis, studies had to meet the following criteria: (1) the focus had to be on the effects of (all types) of martial arts on externalizing behavior, including studies measuring the effects of martial arts on aggression, anger, hostility, anti-social behavior and violence, (2) only studies with a control group consisting of non-athletes or other sport athletes were included, (3) the sample had to consist of children and juveniles up to the age of 19, and (4) statistics suitable for meta-analysis had to be reported.
9
The search yielded 40 relevant studies of which 9 studies (with #ES = 76 effect sizes, N = 1169 participants, and 6 independent non overlapping samples) met the inclusion criteria of this meta-analysis. Appendix A shows the characteristics of the included studies.
Coding the Studies
A detailed coding scheme was used to record all study, design, sample, control group and outcome characteristics that potentially moderate the effects of martial arts on
externalizing behavior. The study characteristics consisted of the impact factor of the journal. An impact factor between 0 and 2 was considered low to moderate and an impact factor higher than 2 was considered high (Stams et al., 2006). For study design characteristics, group assignment (random or not-random) and whether or not martial arts were implemented as an intervention were coded. The coded sample characteristics included age, gender
(proportion of male participants) and information on the martial artists, such as type of martial arts (judo, karate or other martial arts including taekwondo, aikido and kempo), experience in martial arts in years and intensity of martial arts training in hours a week. The control group characteristics included the type of control group (non-athletes or other sport athletes) and the type of sport practiced in the control group (individual sports, team sports or no sports). Finally, for the outcome characteristics the construct category was coded
(aggressive behavior, aggressive attitude or both/other).
Inter-rater agreement was based on a double-coding of all studies by two coders. An inter-rater agreement of 87 percent was found between the two coders.
Data Analysis
In order to establish the overall effects of martial arts on externalizing behavior and to examine which factors moderate this effect, a multi-level meta-analysis was conducted. The outcomes of the studies were transformed into the correlation coefficient effect size (r). Because the variance of the coefficients r depends strongly on the correlation, the coefficients
10 r were recoded into in Fisher’s z values (Lipsey & Wilson, 2001). A positive effect size indicated that martial artists show more externalizing behavior than athletes in other sports or non-athletes, whereas a negative effect size can be interpreted as martial artists showing less externalizing behavior than athletes in other sports or non-athletes. These effect sizes have been calculated using formulas from Wilson (2010), mostly based on means and standard deviations.If the results were reported to be non-significant, without measures to calculate an effect size, an effect size of zero was assigned (Durlak & Lipsey, 1991).
Continuous variables were centered around their mean, and categorical variables with more than two categories were recoded into a series of dummy variables. Extreme values of the effect sizes (>3.29 SD from the mean; Tabachnik & Fidell, 2007) were corrected by winsorizing these outliers. Standard errors and sampling variance of the effect sizes were computed using formulas by Lipsey and Wilson (2001).
Because the data of most studies resulted in more than one effect size, a traditional random effects model was extended to a three-level random effects model (Van den
Noortgate, López-López, Marín-Martínez, & Sánchez-Meca, 2013). The program R was used to conduct multi-level meta-analysis, using a random effects model that allows for the
hierarchical structure of the data. Within this model three form of variations were taken into account: the random sampling variation of observed effect sizes, the variance between outcomes within studies, and the between-study variance. The model was extended by including study, design and sample characteristics as moderators in order to investigate their influence on the relation between participation in martial arts and externalizing behavior problems.
Publication Bias
A common problem in conducting a meta-analysis is that studies with no significant or negative results are less likely to be published than studies with positive and significant
11
results. The studies included in this meta-analysis may therefore not be an adequate representation of all studies that have been conducted. This phenomenon is called the ‘file drawer problem’ (Rosenthal, 1995). In this meta-analysis, the file drawer bias was examined by drawing a funnel plot in MIX 2.0 (Bax, 2011). According to Egger’s method, we will test the asymmetry of the funnel plot (Egger, Davey Smith, Schneider, & Minder, 1997).
Results
The current meta-analysis consists of k = 9 studies (with 6 non overlapping samples), #ES = 76 effect sizes, reporting on N = 1169 juveniles of whom n = 421 participated in martial arts and n = 748 constituting the control group.
Overall Effects
Table 1 presents the results of the multilevel meta-analysis. Cohen’s (1988) guideline was used to interpret the effect sizes, with effect sizes or r = .10 considered small, r = .25 medium and r = .40 large. No significant association was found between participation in martial arts and externalizing behavior (r = -.075, CI = -.244 – .094, p > .05), indicating that there is no overall relation between practicing martial arts and externalizing behavior in juveniles.
When checking for publication bias, Egger’s method did not indicate funnel plot asymmetry, since the regression coefficient was not significant. Furthermore, the trim and fill plot showed that there were no missing studies, which indicates that there is no publication bias (see Figure 1). However, the skewness test was significant (Z = -2.127, p < .05), indicating that the effects size distribution was skewed to the left. This showed that small studies with negative effects were more often reported, suggesting that perhaps a small publication bias was present in the current meta-analysis.
12
Table 1.
Overall and Moderator Effects Relation Martial Arts and Externalizing Behavior
* p <.05, ** p <.01, *** p < .001
Note. # studies = number of independent studies; #ES = number of effect sizes; t0 = difference in mean r with
zero; t1 = difference in mean r with reference category; mean r = mean effect size (r); F(df1, df2) = omnibus test;
(RC) = reference category. # studies # ES β₀ (mean r) t0 β₁ t1 F(df1, df2)
Overall relation martial arts and externalizing behavior
6 76 -.075 -.865 Moderator variables Study characteristics Impact factor 6 76 F(1,74) = .066 Between 0 and 2 (RC) 3 34 -.050 -.364 Higher than 2 3 42 -.099 -.769 -.048 -.256 Design characteristics
Assignment to control group 6 76 F(1,74) = .195
Other (RC) 4 71 -.049 -.439 Random 2 5 -.135 -.827 -.088 -.442 Intervention study 6 76 F(1,74) = .053 No (RC) 4 73 -.088 -.827 Yes 2 3 -.038 -.196 .051 .219 Sample characteristics Gender 6 76 F(1,74) = .077 Male sample (RC) 4 52 -.080 -.888
Female sample or mixed sample 2 24 -.062 -.624 .018 .278
Age 6 76 F(1,74) = 1.392
12 and younger (RC) 2 21 .075 .503
Older than 12 4 55 -.133 -1.399 -.206 -1.180
Type of martial arts 6 76 F(2,73) = 31.925***
Judo (RC) 2 31 .194 3.009**
Karate 2 31 -.022 -.330 -.214 -4.472*** Other 4 14 -.217 -3.297*** -.394 -4.462***
Experience in martial arts 5 F(1,66) = .749
2 years or less (RC) 3 41 .059 .553
More than 2 years 2 35 -.074 -.669 -.133 -.865
Intensity of training in martial arts 4 F(1,63) = 2.338
3 hours a week or less (RC) 2 43 .125 .042
More than 3 hours a week 2 33 .044 1.351 -.081 -1.529
Control group characteristics
Type of control group 6 46 F(1,44) = 26.718***
Non-athletes (RC) 4 23 -.232 -2.120*
Athletes from other sports 3 23 -.002 -.021 .235 5.169***
Type of sport in control group 6 46 F(2,43) = 44.401***
Individual sport (RC) 2 17 .103 1.315 Team sport 2 6 -.224 -2.433* -.319 -3.674*** No sport 4 23 -.172 -2.362* -.271 -6.406*** Outcome characteristics Construct category 6 76 F(2,73) = 4.646 Aggressive behavior (RC) 6 38 -.049 -.536 Aggressive attitude 3 24 -.173 -1.713 -.025 -.349 Both or other 2 14 -.074 -.682 -.124 -2.060*
13 Figure 1. Trim-and-fill plot. Graph from Bax (2011). MIX 2.0 — professional software for meta-analysis in Excel. Version 2.0.1.4. BiostatXL, 2011. Retrieved from: http://www.meta-analysis-madeeasy.com.
The likelihood ratio test comparing models with and without between-study variance (level 3) showed that significant variance was present at the between-study level (σ2level 3 = .034, χ2(1)
= 17.641; p < .0001). The variance between the effect sizes within studies (level 2) was significant as well (σ2
level 2 = .028, χ2(1) = 74.596; p < .0001). About 17% of the total effect
size variance was accounted for by the sample variance (level 1), 38% by the variance between effect sizes within studies (level 2), and 45% by the variance between studies (level 3). Since the variance was significant on the second and third level, moderator analyses of study, design, sample, control group and outcome characteristics were conducted to assess the strength of the relation between martial arts participation and externalizing behavior.
14 Moderator Effects
Study characteristics. The only study characteristic examined in the current
meta-analysis was the impact factor of the journal in which the study was published. This factor did not moderate the relation between martial arts and externalizing behavior (F(1,74) = .066; p = .798). Furthermore, post-hoc analyses with the impact factor included as a continuous moderator showed no moderating effects either.
Design characteristics. The design characteristics included in the moderator analyses
were control group assignment and whether or not martial arts were implemented as an intervention. Both control group assignment (F(1,74) = .195; p = .659) and whether or not an intervention design was applied (F(1,74) = .053; p = .818) were no significant moderators.
Sample characteristics. The type of practiced martial arts had a moderating effect on
the relation between martial arts and externalizing behavior (F(2,73) = 31.925; p < .001; with r = .194 for judo, r = -.022 for karate and r = -.217 for other martial arts). More specifically, judokas tended to show more externalizing behavior compared to the control group, while juveniles practicing other martial arts than judo or karate, showed significantly less
externalizing behavior than the control group. No moderating effects were found for gender (F(1,74) = .077; p = .781), age (F(1,74) = 1.392; p = .238), the experience of the martial artists (F(1,66) = .749; p = .387), and the intensity of the training in martial arts (F(1,63) = 2.338; p = .126). Again, post-hoc analyses with age, experience and intensity included as a continuous moderator showed no moderating effects either.
Control group characteristics. Moderating effects were found for the control group
characteristics. First, the type of control group moderated the effect of martial arts on
externalizing behavior (F(1,44) = 26.718; p < .001; with r = -.232 for non-athletes, versus r = .002 for athletes from other sports). Compared to non-athletes, martial artists showed
15
compared to other sports. Second, the type of sport practiced in the control group had a moderating effect on the relation between participation in martial arts and externalizing behavior in juveniles (F(2,43) = 44.401; p < .001). Martial artists showed significantly less externalizing behavior compared to team sports athletes and non-athletes. There were, however, no differences in externalizing behavior when comparisons were made with individual athletes.
Outcome characteristics. The construct used to measure externalizing behavior (i.e.,
aggressive behavior, attitude, or both), did not moderate the relation between martial arts participation and externalizing behavior (F(2,73) = 4.646; p = .098).
Multivariate Analysis
A multivariate analysis was conducted including all the significant moderating
variables (see Table 2). Before the multivariate analysis, we checked for collinearity between the significant moderators and found that the no sports category of the variable type of sports in the control group and the non-athlete category of the variable type of control group yielded the same effect sizes and thus the same results. Therefore, the no sports category of the variable type of sports in the control group was removed from the multivariate analysis. The multivariate analyses showed that karate as the type of practiced martial arts, a control group of other sports athletes and team sports athletes still yielded unique moderating effects.
16
Table 2.
Results for the Multivariate Model with the Significant Moderators
* p <.05, ** p <.01, *** p < .001
Note. SE = standard error; Z = significance of moderator; F(df1, df2) = omnibus test.
Discussion
The aim of the current meta-analysis was to examine the relation between martial arts participation and externalizing behavior in juveniles. Overall, no relation was found,
suggesting that there was no difference between the level of externalizing behavior for martial artists and athletes in other sports or non-athletes. However, there was variation between the effect sizes within and across studies that could be explained by moderator effects. Therefore, we conducted moderator analyses to examine the influence of study, design, sample, control group and outcome characteristics.
Moderator analyses showed that sample characteristic the type of martial arts influenced the relation between martial arts and externalizing behavior. Judokas showed significantly more externalizing behavior compared to control juveniles, whereas juveniles practicing karate showed no significant differences and juveniles practicing other martial arts showed significantly less externalizing behavior. Furthermore, the relation between
participation in martial arts and externalizing behavior was influenced by control group characteristics. When compared to non-athletes, martial artists showed less externalizing behavior, and when the control group consisted of athletes from other sports, no significant
β (SE) Z F(df1, df2)
Moderator variables
Sample characteristics F(4,41) = 100.342***
Type of martial arts
Karate -.164 (.034) -4.789*** Other -.201 (.182) -1.104
Control group characteristics
Type of control group
Athletes from other sports .277 (.033) 8.361*** Type of sport in control group
17
differences were found. The types of sports practiced in the control group also influenced the relation between martial arts and externalizing behavior. When compared to juveniles
practicing team sports or no sports at all, less externalizing behavior was found in martial artists. Only when compared to individual sports, no significant differences in externalizing behavior were found. No moderating effects were found for the impact factor of the journal, the control group assignment, whether or not martial arts were implemented as an
intervention, age, gender, the amount of experience in martial arts, the intensity of the training in martial arts, and the construct category.
First, we found no overall relation between martial arts and externalizing behavior. This is in line with the results from a previous meta-analysis on the effects of sports on juvenile delinquency, which did not find overall effects on delinquency either (Spruit et al., 2015). Thissuggests that in fact there might not be a relation between any type of sports and externalizing behavior. While contact sports are often associated with aggression, the present study has found that contact sports in existing research have not been well defined through including both team sports and martial arts as contact sports. This appears particularly
relevant, because the current meta-analysis found that team sport athletes specifically seemed to show more externalizing behavior. Future research should therefore aim to better define the construct of contact sports.
The current lack of an overall effect could be due to the small number of studies included in the current meta-analysis. A small number of studies in a meta-analysis decreases statistical power (Begg & Mazumdar, 1994), that is, lower the probability to find a significant overall effect size. It is, however, important to notice that no publication bias was present in this study, which makes it unlikely that unpublished articles with no effects were excluded. Moreover, the multi-level design of the current meta-analysis enabled full exploitation of the available data on the effects of martial arts on externalizing behavior in juveniles.
18
Second, the type of practiced martial arts moderated the effect of martial arts on externalizing behavior. It appears that judokas showed more externalizing behavior than the control group, while karateka showed no effects and other martial arts showed less
externalizing behavior. These outcomes were unexpected, because only traditionally trained martial artists were included in the present study. As traditional martial arts have been associated with a decrease in externalizing behavior (Nosanchuk, 1981), we would only expect to find a difference between traditionally and non-traditionally trained martial artists and not between different traditional martial arts. In addition, one could argue that sports with more direct forms of aggression, including impact sports like karate, would lead to more externalizing behavior in athletes outside the sport context. As judo is the most low-impact sport included in this study, this assumption is not in line with the present results. In contrast, high impact sports could also enable venting off more aggression and externalizing behavior (see ‘catharsis theory’ Breuer & Freud, 1974) than low impact sports. Judo would then enable this ‘venting off of aggression’ the least, which could explain the current outcomes. Finally, previous research has found more anger in children starting judo training (Reynes & Lorant, 2001). The current outcomes could therefore be more indicative of the fact that the judokas in the present study were more aggressive to begin with and this might even have been the reason these juveniles started practicing judo.
Third, the effects of martial arts were dependent on the type of control group (sports) the martial artists were compared to. Martial artists showed less externalizing behavior than non-athletes and team sport athletes, whereas no differences were found between martial artists and individual athletes. The similarity in the levels of externalizing behavior for martial artist and individual athletes can be explained by the fact that martial arts are basically individual sports, and are therefore comparable to other individual sports. Furthermore, a previous meta-analysis on the effects of sports on juvenile delinquency
19
showed that individual sports in general were associated with less delinquency (Spruit et al., 2015), which further explains the findings in the present meta-analysis. The results of this meta-analysis and the results of the previous meta-analysis of Spruit et al. (2015) indicate that possible negative effects of sports participation might not apply to individual sports,
including martial arts.
Finally, in contrast to previous studies that have found that more experience in martial arts led to a decline in externalizing behavior (Lakes & Hoyt, 2004; Skelton, Glynn, & Berta, 1991), the present study found no moderating effect of martial arts training experience. To equally divide the small amount of studies over the different categories, a cut-off point of two years of experience was used. Future research should therefore aim to investigate experience in martial arts with other cut-off points to further distinguish the differences between
beginners and advanced martial artists. However, post-hoc analyses with experience as a continuous moderator still showed no moderating effect, indicating that martial arts training experience in fact has no influence on the relation between martial arts participation and externalizing behavior in juveniles.
Limitations
There are some limitations of this study that need to be addressed. First, only a small number of studies could be included in the present meta-analysis, which is due to the limited amount of research on the effects of martial arts on externalizing behavior in juveniles. Second, the included studies did not always provide enough information about the coded characteristics and some relevant moderators, such as the difference between traditional and modern martial arts (Nosanchuck, 1981) and the coach-athlete relationship (Rutten et al., 2011), could not be examined at all. Furthermore, most studies included a male only sample, so that differences between male and female martial artists regarding externalizing behavior could not be examined. This might explain why gender had no significant moderating effect.
20
Finally, several of the included studies were of low quality, and provided limited information on participants’ assignment to conditions. For these studies, it was unclear to what extent the martial arts and control group were comparable.
Strengths and implications for the future
Despite these limitations, the present meta-analysis has several strengths. A three-level random effects model was used, which increased statistical power and facilitated a more extensive analysis of factors that may moderate the relation between martial arts and
externalizing behavior. Furthermore, the findings of the current meta-analysis yield some important implications for future research and policy. Overall, no relation was found between martial arts and externalizing behavior in juveniles. This finding contradicts existing policies for martial arts participation in juveniles. Some local authorities choose to limit martial arts participation because of the assumed association with violence, whereas others choose to fund martial arts clubs because of the assumed positive effects on juveniles (Van Koolwijk et at., 2014). The present study shows that martial arts may not be useful nor damaging when it comes to decreasing or preventing externalizing behavior, similar to sports in general (Spruit et al., 2015). Moreover, given that sports or other leisure activities only show weak
associations with juvenile problem behavior (e.g. ‘balansmodel’ [balancemodel] Bakker, Bakker, Van Dijke, & Terpstra, 1998), the current lack of association is not really surprising. However, sports, including martial arts, could be combined with other interventions in order to gain positive effects on externalizing behavior. Furthermore, previous research showed that martial arts (and sports in general) may serve other purposes that have not been examined in the present study, such as increasing self-esteem or emotional stability in juveniles (Crews, et al., 2004; Finkenberh 1990; Konzay & Bourdreau, 1984). In addition, martial arts may constitute a context where juveniles with behavior problems may learn how to cope with challenging social problems situations, such as competition, dealing with
21
authorities and perceptions of disadvantage, through targeted treatment (Kirkcaldy, Shephard, & Siefen, 2002).
Conclusion
In sum, the present study showed that there is no overall effect of martial arts on externalizing behavior in juveniles, but our findings reveal that participation in martial arts is related to less externalizing behavior in juveniles compared to non-athletes and team sports athletes. Furthermore, judokas specifically showed more externalizing behavior, whereas participants in other martial arts showed a similar level or less externalizing behavior in juveniles. Future research with more detailed information on sample, sports and control group characteristics is needed to further investigate gender differences and the differences between traditional and modern martial art.
22
References
References marked with an asterisk indicate studies included in this meta-analysis.
Bakker, I., Bakker, C., Van Dijke, A., & Terpstra, L. (1998). O&O in perspectief [O&O in perspective]. Utrecht: NIZW. (In Dutch).
Bax, L. (2011). MIX 2.0 Professional software for meta-analysis in Excel: Version 2.0.1.4. BiostatXL. Retrieved from http://www.meta-analysis-made-easy.com.
Begg, C. B., & Mazumdar, M. (1994). Operating characteristics of a rank correlation test for publication bias. Biometrics, 50, 1088-1101.
Bredemeier, B. J., Shields. D. L., Weiss. M. R., & Cooper, B. A. B. (1986) The relationship of sport involvement with children's moral reasoning and aggression tendencies. Journal of Sport Psychology, 8, 304-318.
Breedveld, K., Kamphuis, C., & Tiessen-Raaphorst, A. (2008). Rapportage Sport 2008 [Report Sport 2008]. Den Haag: Sociaal en Cultureel Planbureau (SCP)/ W.J.H. Mulier Instituut. (In Dutch).
Breuer, J., Freud, S. (1974). Studies on hysteria. Harmondsworth: Penguin Books.
Burton, J. M., & Marshall, L. A. (2005). Protective factors for juveniles considered at risk of criminal behavior: does participation in extracurricular activities help? Criminal Behavior and Mental Health, 15, 46-64.
Cohen, J. (1988). Statistical power analysis for the behavioral science. Hillsdale, NJ: Lawrence Erlbaum.
Conroy, D. E., Silva, J. M., Newcomer, R. R., Walker, B. W., & Johnson, M. S. (2001). Personal and participatory influences on the socialization of aggressive sport behavior. Aggressive Behavior, 27, 405-418.
23
psychological well-being in low-income hipanic children. Perceptual and Motor Skills, 98, 319-324.
Daniels, E., & Leaper, C. (2006). A longitudinal investigation of sport participation, peer acceptance, and self-esteem among adolescent girls and boys. Sex Roles, 11-12, 875-880.
Daniels, K. & Thornton, E. (1990). An analysis of the relationship between hostility and training in the martial arts. Journal of Sports Sciences, 8, 95-101.
Daniels, K. & Thornton, E. (1992). Length of training, hostility and the martial arts: a comparison with other sporting groups. British Journal of Sports Medicine, 26, 118-120.
Danish, S., Kleiber, D., & Hall, H. (1987). Developmental intervention and motivation enhancement in the context of sport. In M. L. Maehr, & D. A. Kleiber, Advances in motivation and achievement: A research annual (p. 211-235). Greenwich, DT: JAI Press
Danish, S. J., Nellen, V., & Owens, S. (1996). Community-based life skills programs: Using sports to teach life skills to adolescents. In J. L. Van Raalte & B. W. Brewer,
Exploring sport and exercise psychology (p. 205-225). Washington, DC: American Psychological Association.
Danish, S. J., Petitpas, A., & Hale, B. (1995). Psychological interventions: A life
development model. In S. M. Murphy, Sport psychology interventions (p. 19-28). Champaign, IL; Human Kinetics Publishers.
De Knop, P., Engström, L. M., Skirstad, B., & Weiss, M. (1996). Worldwide Trends in Juveniles Sport. Champaign: Human Kinetics Publishers.
*Delva-Tauiliili, J. (1995). Does brief aikido training reduce aggression of juveniles? Perceptual and Motor Skills, 80, 297-298.
24
Durlak, J. A., & Lipsey, M. W. (1991). A practitioner's guide to meta-analysis. American Journal of Community Psychology, 19, 291–332.
Edelman, A. J. (1994). The implementation of a video-enhanced aikidobased school violence prevention training program to reduce disruptive and assaultive behavior among secerely emotionally disturbed adolescents. Ed: Practicum, D. Nova Southeastern University. ERIC Document Reproduction Service No. EC 3040227.
Egger, M., Davey Smith, G., Schneider, M., & Minder, C. E. (1997). Bias in meta-analysis detected by a simple, graphical test. British Medical Journal, 315(7129), 629-634. Endresen, I. M., & Olweus, D. (2005). Participation in power sports and antisocial
involvement in preadolescent and adolescent boys. Journal of Child Psychology and Psychiatry. 46, 468-478.
Finkenberg, M. (1990). Effect of participation in taekwondo on college women’s self- concept. Perceptual and Motor Skills, 17, 891-894.
Graczyk, M., Hucinski, T., Norkowski, H., Peczak-Graczyk, A., & Rozanowska, A. (2010). The level of aggression syndrome and a type of practised combat sport. Journal of Combat Sport and Martial Arts, 1, 1-14.
Hansen, D., Larson, R., & Dworkin, J. (2003).What adolescents learn in organized juveniles activities: A survey of self-reported developmental experiences. Journal of Research on Adolescence, 13(1), 25-56.
Hedstrom, R., & Gould, D. (2004). Research in juveniles sports: Critical issues status. East Lansing, MI: Institute for the Study of Juveniles Sports.
Kellogg, S. (2013). The impact of martial arts training on adolescents (Master’s thesis). Texas Tech University, Lubbock, United States.
25
activity and self-image and problem behaviour among adolescents. Social Psychiatry and Psychiatric Epidemiology, 37, 544-550.
Konzak, B., & Boudreau, F. (1984). Martial arts training and mental health: An exercise in self-help. Canada’s Mental Health, 32, 2-8.
Kurian, M., Caterino, L. C. & Kulhavy, R. W. (1993). Personality characteristics and duration of ATA tae kwon do training. Perceptual and Motor Skills, 76, 363-366. Lakes, K. D., & Hoyt, W. T. (2004). Promoting self-regulation trough school-based marital
arts training. Applied Developmental Psychology, 25, 283-302.
Lamarre, B. W., & Nosanchuk, T. A. (1999). Judo - the gentle way: A replication of studies on martial arts and aggression. Perceptual and Motor Skills, 88, 992-996.
Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: Sage *Lotfian, S., Ziaee, V., Amini, H., & Mansournia, M. A. (2011). An analysis of anger in
adolescent girls who practice the martial arts. International Journal of Pediatrics, 2, 1-5.
Marsh, H. W., & Kleitman, S. (2003). School athletic participation: Mostly gain with little pain. Journal of Sport and Exercise Psychology, 25, 205-228.
Nosanchuk, T. A. (1981). The way of the warrior: The effects of traditional martial arts training on aggressiveness. Human Relations, 34, 435-444.
Nosanchuk, T., & MacNeil, M. (1989). Examination of the effects of traditional and modern martial arts training on aggressiveness. Aggressive Behavior, 15, 153-159.
Papacharisis, V., Goudas, M., Danish, S. J., & Theodorakis, Y. (2005). The effectiveness of teaching a life skills program in a sport context. Journal of Applied Sport Psychology, 17(3), 247-254.
26
al. (2006) .Externalizing and oppositional behaviors and karate-do: the way of crime prevention: A pilot study. International Journal of Offender Therapy and
Comparative Criminology, 50, 654-660.
Reynes, E., & Lorant, J. (2001). Do competitive martial arts attract aggressive children? Perceptual and Motor Skills, 93, 382-386.
*Reynes, E., & Lorant, J. (2002a). Effect of traditional judo training on aggressiveness among young boys. Perceptual and Motor Skills, 94, 21-25.
*Reynes, E., & Lorant, J. (2002b). Karate and aggressiveness among eight-year-old boys. Perceptual and Motor Skills, 94, 1041-1042.
*Reynes, E., & Lorant, J. (2004). Competitive martial arts and aggressiveness: a 2-yr. longitudinal study among young boys. Perceptual and Motor Skills, 98, 103-115. *Rutten, E. A., Schuengel, C., Dirks, E., Stams, G. J. J. M., Biesta, G. J. J., & Hoeksma, J. B.
(2011). Predictors of antisocial and prosocial behavior in an adolescent sports context. Social development, 20(2), 294-315.
Rosenthal, R. (1995). Writing meta-analytic reviews. Psychological Bulletin, 118, 183-192. Scanlan, T. K., Babkes, M. L., & Scanlan, L. A. (2005). Participation in sport: A
developmental glimpse at emotion. In J. L. Mahoney, R. W. Larson, & J. S. Eccles, Organized activities as contexts of development. Hillsdale, NJ: Lawrence Erlbaum. Skelton, D. L., Glynn, M. A., & Berta, S. M. (1991). Aggressive behavior as a function of
taekwondo ranking. Perceptual and Motor Skills, 72, 179-182.
Smoll, F., & Smith, R. (1996). Children and juveniles in sport: A biopsychosocial perspective. Chicago: McGraw-Hill.
Spruit, A., Van Vugt, E. S., Van der Put, C. E., Van der Stouwe, T., & Stams, G. J. J. M. (2015). Sport participation and juvenile deliquency: A meta-analytic review. Manuscript submitted for publication.
27
Stams, G. J. J. M., Brugman, D., Decovic, M., Van Rosmalen, L., Van der Laan, P. H., & Gibbs, J. C. (2006). The moral judgment of juvenile delinquents: A meta-analysis. Journal of Abnormal Child Psychology, 34, 697–671.
*Steyn, B., & Roux, S. (2009). Aggression and psychological well-being of adolescent taekwondo participants in comparison with hockey participants and non-sport group. African Journal for Physical, Health Education, Recreation and Dance, 15, 32-43. Tabachnik, B. G., & Fidell, L. S. (2007). Using multivariate statistics. Boston: Allyn &
Bacon.
Tammelin, T., Näyhä, S., Hills, A., & Järvelin, M. (2003). Adolescent Participation in Sports and Adult Physical Activity. American Journal of Preventive Medicine, 24, 22-28. Trulson, M. E. (1986). Martial arts training. A novel cure for juvenile delinquency. Human
Relations, 39, 1131-1140.
Ullrich-French, S., McDonough M. H., & Smith A. L. (2012) Social connection and psychological outcomes in a physical activity-based juveniles development setting. Research Quarterly for Exercise and Sport, 83, 431–441.
Van Koolwijk, M., Langendijk, E., & Van Egdom, H. (2014). Eindrapportage regulering full contact vechtsporten: Regulering vechtsport is samenspel [Final report regulation full contact martial arts: Martial arts regulation is teamwork]. Arnhem: NOC*NSF. (In Dutch).
Van den Noortgate, W., López-López, J. A., Marín-Martínez, F., & Sánchez-Meca, J. (2013). Three-level meta-analysis of dependent effect sizes. Behavior Research Methods, 45, 576-594.
Wilson, D. B. (2010). Practical meta-analysis effect size calculator. Retrieved, 2015, from.
http://gemini.gmu.edu/cebcp/EffectSizeCalculator/
28
adolescent boy athletes: A comparison among judo, karate, swimming and non-athletes. Iranian Journal of Pediatrics, 22(1), 9-14.
*Zivin, G., Hassan, N., DePaula, G., Monti, D., Harlan, C., Hossain, K., et al. (2001). An effective approach to violence prevention: traditional martial arts in middle school. Adolescence, 36, 443- 459.
29 Appendix A
Characteristics of Included Studies
Study Study characteristics Design characteristics Sample characteristics Secondary article Tertiary article Year N # r (M) Impact- factor Assignment Intervention study % male
Age Experience Intensity training Lotfian Ziaee (2012) - 2011 291(f), 201(m) 32 (-.047) 3.736 Other No 50 >12 >2 >3 Delva-Tauiliili - - 1995 42(m) 1 (-.097) .658 Other Yes 100 ≤12 ≤2 >3 Reynes Reynes (2002) Reynes (2004) 2002 64(m) 30 (.126) .658 Other No 100 ≤12 ≤2 ≤3 Zivin - - 2001 60(m) 2 (0) 2.312 Random Yes 100 >12 ≤2 ≤3 Rutten - - 2011 439(m) 3 (-.251) 1.382 Random No 100 >12 >2 - Steyn - - 2009 41(f), 31(m) 8 (-.340) 4.030 Other No 43.1 >12 - -
Sample characteristics Control group characteristics Outcome characteristics Type of martial arts (# r) Type of control group (# r) Type of sports practiced in control group (# r) Construct category (# r) Judo Karate Other martial arts Non-athletes Other sport athletes Individual sports Team sports No sport AB AA B/O Lotfian 16 16 0 16 16 16 0 16 8 16 8 Delva-Tauiliili 0 0 1 1 0 0 0 1 1 0 0 Reynes 15 15 0 - - - 18 6 6 Zivin 0 0 2 2 0 0 0 2 2 0 0 Rutten 0 0 3 0 3 1 2 0 3 0 0 Steyn 0 0 8 4 4 0 4 4 6 2 0
Note. N = number of participants; # r (M) = number of effect sizes (mean); impact factor = impact factor of journal in 2013; Assignment = how participants were assigned to a control group; Intervention study = martial arts are implanted as an intervention yes/no; % male = percentage of males in sample; Age = mean age of sample; Experience = the experience of martial artists in years; Intensity training = the intensity of training martial arts in hours a week; f = number of female participants; m = number of male participants; Random = random assignment to control group; Other = other way of assignment to control group (not random); >12 = mean age above12 years old; ≤12 = mean age of 12 years old or below 12 years old; >2 = more than 2 years of experience; ≤2 = 2 years of experience or less; >3 = more than 3 hours of training a week; ≤3 = 3 hours of training a week or less; AB = aggressive behavior; AA = aggressive attitude; B/O = both or other.
