A Systematic Review of Life Skill Development Through Sports Programs Serving Socially Vulnerable Youth

ABSTRACT Purpose: Despite the strong belief in sports programs as a setting in which socially vulnerable youth can develop life skills, no overview exists of life skill development in sports programs serving this youth group. Therefore, the present systematic review provides an overview of the evidence on life skill development in sports programs serving socially vulnerable youth and, insofar as it was investigated in the included studies, of the conditions conducive to life skill development in these sports programs. Method: Potentially relevant studies published during 1990 to 2014 were identified by a search in 7 electronic databases. The search combined terms relating to (a) sport, (b) youth AND socially vulnerable, and (c) life skills. Eighteen of the 2,076 unique studies met the inclusion criteria. Results: Each included study reported that at least 1 life skill improved in youth who participated in the studied sports program. Improvements in cognitive and social life skills were more frequently reported than were improvements in emotional life skills. Only a few of the included studies investigated the conditions in the studied sports programs that made these programs conducive to life skill development. Conclusions: Sports programs have the potential to make a difference in the life skill development of socially vulnerable youth. This conclusion needs to be treated with some caution, because the studies experienced many challenges in reducing the risk for bias. Several alternative research strategies are suggested for future studies in this field.

This systematic review provides an overview of the evidence on life skill development in sports programs serving socially vulnerable youth. Socially vulnerable youth represent a broad group, ranging from youth living in areas of low socioeconomic status (SES) to youth receiving residential care or nonresidential counseling. A common denominator is that they face stressors in their everyday life, such as income poverty, poor family management, low housing quality, and peers being involved in problem behavior (Feinberg, Jones, Greenberg, Osgood, & Bontempo, 2010;Hawkins, Catalano, & Arthur, 2002). These stressors may lead to feelings of incompetence, social disconnectedness, negative experiences with societal institutions (e.g., family, school, and health care), a lack of ambition (Haudenhuyse, Theeboom, & Skille, 2014;Vettenburg, 1998), and a reduced chance for participating in meaningful activities, such as sports (Vandermeerschen, Vos, & Scheerder, 2013). Programs aiming to support youth in dealing with stressors of everyday life are often grounded in the positive youth development (PYD) perspective, which emphasizes that youth, including those who are socially vulnerable, have the potential to develop the life skills they need to deal with the stressors they face (Damon, 2004;Lerner et al., 2005). Moreover, life skills are important predictors of youths' future well-being, academic performance, and job satisfaction (Ridder, Lensvelt-Mulders, Hinkemeyer, Stok, & Baumeister, 2012;Zins, Weissberg, Wang, & Walberg, 2004).
Life skills, defined as "those skills that enable individuals to succeed in the different environments in which they live, such as school, home and in their neighborhoods" (Danish, Forneris, Hodge, & Heke, 2004, p. 40), can be divided into emotional, cognitive, and social skills (Lerner et al., 2005;Turnnidge, Côté, & Hancock, 2014). Emotional life skills pertain to one's internal sense of well-being and self-worth (Lerner et al., 2005). The development of emotional skills is particularly important for socially vulnerable youth because they face mental health problems more often than do nonvulnerable youth (Reiss, 2013;Wille, Bettge, & Ravens-Sieberer, 2008). Cognitive life skills pertain to abilities such as self-regulation, decision making, and impulse control (Lerner et al., 2005). Such cognitive skills are shown to be protective factors for the stressors that socially vulnerable youth face in their everyday lives (Lösel & Farrington, 2012). Social skills pertain to skills that can be used in interpersonal relationships, such as communication skills, conflict resolution, and prosocial behavior (Lerner et al., 2005). The development of social skills is important for socially vulnerable youth because those skills may help them to decrease social disconnectedness, which is one of the major indicators for social vulnerability (Haudenhuyse et al., 2014).
Sports programs (i.e., formally structured activities that take place for a specific time period and in the presence, or under the instruction, of adults (E. Bean, Whitley, & Gould, 2014;Fuller, Percy, Bruening, & Cotrufo, 2013;Ullrich-French, McDonough, & Smith, 2012)), are believed to provide youths with settings for life skill development (Fraser-Thomas, Côté, & Deakin, 2005). It has been argued that positive experiences in sports programs lead to improved emotional life skills, such as increased self-worth or reduced depressive symptoms (Eime, Young, Harvey, Charity, & Payne, 2013). A suggested mechanism for the development of cognitive life skills is the goal-setting behavior required in the sports setting, which gives young people the opportunity to hone their cognitive skills (Jonker, Elferink-Gemser, & Visscher, 2011). Moreover, the sports setting is an environment rich in feedback, which is considered to be a prerequisite for the development of cognitive life skills such as self-regulation skills (Jonker et al., 2011). Besides these social and psychological mechanisms, physiological mechanisms are suggested for the relationship between sports participation and emotional and cognitive outcomes. For instance, it has been shown that physical activity leads to changes in neurotransmitters associated with improved well-being (Lubans, Plotnikoff, & Lubans, 2012), self-esteem (Cerin, 2010), and executive functioning (Diamonds & Lee, 2011). Finally, Bailey, Hillman, Arent, and Petitpas (2013) suggested that as many sports programs take place in a social setting, such programs provide youth with opportunities to develop social skills such as communication skills, conflict resolution, and empathy.
To develop and run sports programs serving socially vulnerable youth, policymakers and social professionals will benefit from an overview of what is known about life skill development in such programs. In a systematic review of the psychological and social benefits of participation in sport for children and adolescents, Eime et al. (2013) found that improved self-esteem, social interaction, and fewer depressive symptoms were the most commonly reported psychological and social benefits of sports participation. However, the possibility of generalizing their findings to socially vulnerable youth is limited, because their overview did not distinguish between socially vulnerable and nonvulnerable youth. Such a distinction is needed because the mechanisms underlying life skill development through sports may differ for these two youth groups (Haudenhuyse et al., 2014). Also, Eime et al. did not distinguish between studies on sports programs and so-called sports-fordevelopment programs or sport-plus programs. Sports-for-development programs are sports programs intentionally structured to serve socially vulnerable youth's sports participation and/or life skill development (Coalter, 2015). Only Lubans et al. (2012) systematically reviewed studies on the benefits of sports programs serving socially vulnerable youth. The authors provided an overview of quantitative studies published during 1990 to 2011 on the impact of outdoor activity programs, sports and skill-based programs, and physical fitness programs (e.g., aerobics and circuit training) on the social and emotional wellbeing of socially vulnerable children and youth aged 4 to 18 years old. Based on six studies that examined the benefits of sports programs, Lubans et al. concluded that sports programs potentially have beneficial outcomes for socially vulnerable youth but stated that their findings should be treated with caution because of the low number of included studies.
It is useful to expand on Lubans et al.'s (2012) work with more recent studies for several reasons. First, to our knowledge, additional quantitative studies that have examined life skill developments in sports programs serving socially vulnerable youth have been published since Lubans et al.'s review (e.g., D'Andrea, Bergholz, Fortunato, & Spinazzola, 2013;and Terry, Hahn, & Simjanovic, 2014). Second, it is useful to expand on Lubans et al.'s work with qualitative studies that have described sports coaches', parents', and youth's perceptions of life skill development in these programs, such as the studies by Beaulac, Kristjansson, and Calhoun (2011) and Riley and Anderson-Butcher (2012). A final reason for conducting the present review is that previous reviews in this field have ignored the conditions conducive to life skill development that have been investigated in the included studies. This is unfortunate, because it has been shown that sports programs need to meet certain conditions to provide a setting that supports life skill development (Coalter, 2015;Fraser-Thomas et al., 2005), such as positive peer relationships and sports coaches creating a task-oriented sports climate (Haudenhuyse et al., 2014;Newton et al., 2007;Smith, Smoll, & Cumming, 2007). Conducive conditions seem to be particularly important in sports programs serving socially vulnerable youth. For instance, the emotional baggage of these youth may lead to negative sport experiences and mechanisms of exclusion in sports settings that emphasize competition and masculinity (C. F. Bean, Fortier, Post, & Chima, 2014;Haudenhuyse et al., 2014). Such exclusion in sports and negative sports experiences have been found to increase feelings of rejection and social isolation that can further push these youth down the spiral of vulnerability (Super, Wentink, Verkooijen, & Koelen, 2017).

Study aim
To summarize, despite the increased attention on sports programs as a setting for life skill development in socially vulnerable youth, no recent overview exists of quantitative and qualitative studies investigating life skill development in sports programs serving this group. Therefore, the main aim of the present review was to describe the evidence of life skill development in sports programs serving socially vulnerable youth from quantitative and qualitative studies. As previous systematic reviews have not addressed conditions conducive to life skill development that have been investigated in studies on life skill development in sports programs, an additional aim of this review was to describe what is known about conducive conditions from studies on life skill development in sports programs serving socially vulnerable youth. Accordingly, two research questions were formulated: (a) What is the evidence of life skill development in sports programs serving socially vulnerable youth from both quantitative and qualitative studies? And (b) what is known about conducive conditions for life skill development from these studied sports programs? By addressing these two research questions, this review supports the knowledge base that will help policymakers and practitioners to select and develop sports programs serving socially vulnerable youth and that will support researchers to develop new studies in this field.

Methods
To identify relevant studies, the first author developed a search in continuous deliberation with the other authors. The search combined terms relating to (a) sport, (b) youth AND socially vulnerable, and (c) life skills. In line with Turnnidge et al. (2014), the major outcome variables were terms related to emotional, cognitive, and social life skills. The full search is available in the Appendix. We did not add search terms for Research Question 2, because we aimed to explore what is known about conducive conditions for life skill development in studies included to answer Research Question 1. The search was carried out in seven electronic databases (Scopus, SportDiscus, PsycINFO, SOCindex, Psychology and Behavioral Sciences Collection, PubMed, and Web of Science), resulting in a set of 2,674 records. After removing 598 duplications, the first and second author screened the titles and abstracts of 2,076 unique studies published during 1990 to December 31, 2014 (see Figure 1). This article has no statement of ethical approval or informed consent because it is a study review.
To be included, studies had to meet three initial criteria. The study population had to be socially vulnerable youth aged 10 to 23 years, the setting had to be a sports program serving this youth group, and life skill development had to be reported. Studies were excluded if they did not present primary data, were not published in English or Dutch, and/or were published before 1990. Both authors started to screen titles and abstracts independently. After 1 day of screening, when each author had screened the titles and abstracts of 67 studies, they compared their decisions. The interrater reliability of the decisions between the two authors was strong: 0.96 (McHugh, 2012). Having discussed the 3 studies on which they decided differently, the authors agreed to exclude studies at this screening stage only if it was obvious that the study did not meet the inclusion criteria. The remaining articles were divided between the two authors to screen titles and abstracts, and 147 studies remained for full-text assessment after this screening process.
The first and second authors read the full texts of all 147 studies and independently considered the inclusion of the studies. At this stage, two exclusion criteria were added. We excluded studies if the sports activity was not the core element of the program but merely one of the program elements among several other nonsports elements. We also excluded studies if they only reported on outcomes directly related to the sports context, such as physical activity efficacy or teamwork in the sports setting. If the authors disagreed on a study, the third author read the study and discussed with the first and second author whether it should be included or excluded. After the full-text assessment, 22 studies remained.
The last inclusion criterion was that the study had to be of medium or high rigor. To exclude studies of low rigor, we assessed the rigor of the 22 studies using 10 criteria derived from the TAPUPAS (Transparency, Accuracy, Purposivity, Utility, Propriety, Accessibility, Specificity) framework (Pawson, Boaz, Grayson, Long, & Barnes, 2003). The TAPUPAS framework was chosen to develop the rigor assessment criteria, because it contains general quality criteria that can be applied to both quantitative and qualitative studies. The 10 criteria were: clear description of study aim; appropriate size of sample; sound selection/sampling of sample; appropriate description of the context of the study and of the study participants; conclusions supported by the data; sound description of limitations; sound data; appropriate analysis to answer the research question; logical, traceable, and clear documentation of the research process; and sound extrapolation of conclusions to the theoretical population. To assess the rigor of the studies, the first two authors independently scored the studies on each of the 10 criteria: 1 point if a study satisfied a criterion and 0 points if a study did not satisfy a criterion. Hence, in total, studies could be assigned 10 points, or 1 for each criterion. When the first and second author disagreed about the rigor assessment of a study, they discussed it with the third author, who also read all 22 studies. Studies that received fewer than 5 points were assigned low rigor, studies that received 5 points to 7 points were assigned medium rigor, and studies that received 8 or more points were assigned high rigor (Van Dillen, Van Binsbergen, Koelen, & Hiddink, 2013). At the end of this stage, we excluded 4 studies because they were of low rigor, leaving 18 studies in the synthesis.
To synthesize the data, the first author extracted data from the included studies regarding the type of sports program, the participants, the study design and methods, and the study results. Thereafter, the first author wrote summaries of the extracted data and had ongoing discussions on the synthesis of the data with the second author. The four authors discussed the summaries and synthesis several times in the review process.

Results
This section starts with an overview of the studied sports programs and a brief summary of their designs and methods. Thereafter, the findings of the studies are presented along the three major outcomes: emotional life skills, cognitive life skills, and social life skills. Finally, we give an overview of what the studied sports programs tell us about conditions conducive for life skill development. Table 1 gives an overview of the 18 sports programs studied. The programs were conducted in urban areas in six countries, but mostly in the United States (n = 10) or Canada (n = 3). The settings in which they were conducted were schools in low-SES areas (n = 9), summer camps serving socially vulnerable youth (n = 4), community centers or community sports clubs in deprived areas (n = 3), and residential care (n = 2). In most of the programs, youth participated voluntarily. In some of the programs in the school setting, youth were selected for participation by a school staff member. Twelve of the 18 programs included PYD principles (see Table 1). In line with PYD theory (Damon, 2004), these programs aimed to foster positive development. Some of these programs included PYD principles by building upon the teaching personal and social responsibility (TPSR) model, which is based on the idea that youth can develop effort and teamwork skills, self-direction and goal setting, leadership and helping skills, and respect for others if they play an active role in coaching the sports activities (Hellison, 2003). Other PYD principles included in the sports programs were a mastery sports climate (e.g., Beaulac et al., 2011), a positive peer and coach-youth relationship (e.g., E. Ullrich-French et al., 2012), and facilitation of positive sports experiences (e.g., Holt, Sehn, Spence, Newton, & Ball, 2012;Terry et al., 2014). In addition to the sports programs based on PYD principles, four programs included principles that should empower youth to engage in program development and in decisionmaking processes in the sports activities. For instance, the program studied by Bruening, Dover, and Clark (2009) was based on theories of engagement and freechoice learning, and the program studied by Bonhauser et al. (2005) involved youth in deciding which sports were offered in the sports program. Finally, one program was based on therapeutic procedures (D' Andrea et al., 2013), and one study did not mention the program's theory base (Hasanpour, Tabatabaei, Alavi, & Zolaktaf, 2014). The six programs that were not based on PYD principles aimed to increase physical activity and physical fitness, rather than foster positive developments (see Table 1).

Designs and methods
The study designs and methods varied greatly between studies and ranged from experimental quantitative studies to qualitative interview studies (see Table 2). Most of the quantitative studies applied validated instruments to assess youth's life skills, but different instruments were used to assess the same life skill across studies. For instance, to assess self-esteem, Bonhauser et al. (2005) used the Tennessee Self-Concept Scale; Laberge, Bush, and Chagnon (2012) used the Rosenberg Global Self-Esteem Scale; and Hasanpour et al. (2014) used the Coppersmith Self-Esteem Inventory. In addition to youth, the participants in the qualitative studies involved parents and program staff members.

Emotional life skills
Six of the included studies reported on emotional life skills, and four of them reported improvements (see Table 3). Overall, the findings presented in this section give some indications that sports programs serving socially vulnerable youth are settings in which internalizing symptoms decrease. However, the findings regarding the development of other emotional life skills (e.g., global self-worth, mood, and hope) were mixed. An interesting observation was that improvements in emotional life skills were more frequently reported in quantitative studies than in qualitative studies. Decreases in internalizing symptoms were reported in two quantitative studies (Bonhauser et al., 2005;D'Andrea et al., 2013), which were conducted in two very different settings. D' Andrea et al. (2013) assessed whether participation in basketball activities in addition to treatment as usual affected internalizing symptoms (i.e., depression, anxiety, withdrawal, and somatic complaints) in traumatized girls in a residential treatment center. The authors found that internalizing symptoms in the girls who participated in the basketball activities decreased, as compared with a comparison group of girls who received treatment as usual (D'Andrea et al., 2013). Bonhauser et al. (2005) reported that anxiety Boys and girls aged 9-11 years Voluntary participation of youngsters from a school in a lowincome and minority neighborhood One hour of basketball practice every week for a period of 2 school years. The first eight sessions were used to establish norms, from the ninth session focus on selfdirection, goal setting, leadership, and helping. Games led by the participants themselves.
Note. TPSR = teaching personal and social responsibility. symptoms, but not depressive symptoms, reduced for youth who participated in weekly sports sessions at secondary schools in a deprived area.
The studies that included data on other emotional life skills provide a mixed picture. First, while Beaulac et al. (2011) reported an improved mood in youth participating in weekly dance classes at a community recreation center, Terry et al. (2014) reported no change in mood in youth participating in 19 boxing sessions at schools. A possible explanation for this difference in findings is the difference in sports settings.  (2001) did not. An explanation for these mixed results may be that the sports coaches in the summer camp studied by Ullrich-French et al. received training on how to provide a supportive atmosphere and positive coach-youth connection, whereas the sports coaches in the program studied by Bonnette et al. did not seem to pay specific attention to positive youth-coach relationships. Ullrich-French et al. also assessed developments in hope, defined as belief in the ability to find routes to goals, but they found it did not improve.

Cognitive life skills
Eleven studies reported on the development of cognitive life skills, which can be divided into two categories: self-regulation skills and self-esteem/self-confidence (see Table 3). Overall, each study that reported on such skills revealed that at least one cognitive life skill improved. Qualitative and quantitative studies were equally represented in the 11 studies.
The self-regulation skills that were reported as improving were very diverse (i.e., self-motivation, effort, future focus, goal setting, self-direction, critical thinking, self-concept, self-efficacy, resistance skills, concentration/attention, self-control, taking responsibility for one's own actions, and discipline). The settings of the sports programs in which youth were reported as developing self-regulation were diverse as well. For instance, one program included competitive sports activities at a community sports club (E. , whereas another program was a sportsbased summer camp in which program staff tried to create a mastery-oriented environment (Ullrich-French et al., 2012). Four of the programs in which participants were reported as developing self-regulation skills were based on PYD principles (E. Fuller et al., 2013;Riley & Anderson-Butcher, 2012;Walsh, Ozaeta, & Wright, 2010). One example of a PYD-based sports program that strengthened the development of self-regulation (i.e., self-efficacy, self-concept, and resistance skills) was a program that tried to create a supportive and empowering environment (Fuller et al., 2013). Other examples were a program in which sports coaches encouraged and supported youth to coach the sports activities (Walsh et al., 2010), a program in which coaches conducted a mastery-oriented coaching style (E. , and a program that focused on respect, effort, self-direction, and caring in the sports program context (Riley & Anderson-Butcher, 2012). From interviews with youth, their parents, and program staff, E. , Anderson-Butcher (2012), andWalsh et al. (2010) reported that they were settings where youth developed self-regulation skills like discipline, initiative taking, effort, self-direction, goal setting, self-motivation, and future focus. An interesting observation was that most studies that showed improvements in self-regulation were qualitative studies. The two studies that quantitatively assessed self-regulation skills revealed improvements in attention/concentration (Laberge et al., 2012) and critical thinking (Bonnette et al., 2001). In contrast to these positive results, self-control was not found to increase in the study by Laberge et al. (2012). Improvements in self-esteem and self-confidence were reported in multiple studies. These studies were again very diverse in terms of the setting of the sports programs and the research methods. For instance, improvements in self-esteem were reported in a randomized controlled trial on the impact of aerobic sessions on orphan girls' self-esteem (Hasanpour et al., 2014), in a quasiexperimental study on the impact of weekly sports sessions at schools in deprived areas (Bonhauser et al., 2005), in interviews with parents of youth who participated in a sports-based summer camp (Riley & Anderson-Butcher, 2012), and in pupil profiles written by teachers (Armour & Duncombe, 2012). Selfconfidence, defined as a willingness to try new things, was reported as improving in disengaged youth selected by school teachers to participate in a 1-week outdoor sports activities program (Armour & Sandford, 2013) and in youth who participated in weekly dance classes at a community recreation center (Beaulac et al., 2011). On the basis of interviews with parents and program staff, Beaulac et al. (2011) offered a possible explanation for growth in self-confidence, which was that the dance classes were a setting in which youth could experience feelings of success, which in turn may have increased their confidence in trying new activities in nonsports settings. In contrast to these positive findings on self-esteem and self-confidence, Laberge et al. (2012) did not find that self-esteem improved in a quasiexperimental quantitative study on 16 weeks of lunchtime sports sessions at schools in deprived areas. The authors suggested that self-selection bias (i.e., youth high in self-esteem were more likely to participate than youth low in self-esteem) may be a reason for why they did not find an increase in self-esteem (Laberge et al., 2012).

Social life skills
Fourteen studies reported on developments in social life skills, and 12 of them showed improvements.
Most of these reported social life skills can be divided into two broad categories (i.e., social responsibility skills and social interaction skills). Besides the two broad categories, several other social skills were examined in the studied sports programs. An interesting observation was that all 7 qualitative studies reported improvements in social life skills, whereas only 5 of 11 quantitative studies did. Improvements in social responsibility skills were reported in seven studies. Five of these programs were based on the previously mentioned TPSR model, which aims to develop personal and social responsibility. The authors suggested several possible explanations for why  Communication skills + E.  Conflict resolution skills + Fuller et al. (2013) Conflict resolution skills + Riley & Anderson-Butcher (2012) Conflict resolution skills + participation in the sports program could have led to improvements in social responsibility, and many of these explanations refer to elements of the TPSR model. One example of such an explanation was that the sports coaches continuously discussed the transference of self-direction and goal setting from the sports setting to other settings (Anderson-Butcher, Riley, Amorose, Iachini, & Wade-Mdivanian, 2014;Walsh et al., 2010). Other authors suggested that improvements in social responsibility may be explained by the involvement of caring sports coaches (Hellison & Wright, 2003), the inclusion of a life skill education part (Riley & Anderson-Butcher, 2012), and the presence of negative events, like injuries, that provide opportunities to develop empathy (Holt et al., 2012). Besides studies on sports programs based on the TPSR model, two other studies showed improvements in social responsibility skills. A possible mechanism suggested for the improvement experienced in these sports programs was that the sports program may be a setting in which youth notice that different people have different competencies, which in turn may increase respect for diversity (Beaulac et al., 2011). Armour and Sandford (2013) did not provide an explanation for the improvement in empathy. Two social interaction skills (i.e., communications skills [Armour & Sandford, 2013;Fuller et al., 2013;Riley & Anderson-Butcher, 2012] and conflict resolution skills [E. Fuller et al., 2013;Riley & Anderson-Butcher, 2012]) were reported as improving. An interesting observation is that the four studies that reported improvements in social interaction skills were qualitative studies.
In addition to the studies on social responsibility and social interaction, five studies reported on developments in other social life skills, with mixed results (see Table 3). A positive finding was that social skills improved in youth who participated in weekly sports activities at schools in a deprived area (Armour & Duncombe, 2012). Laberge et al. (2012) and Ullrich-French et al. (2012), however, provided a mixed picture regarding developments in social competence. A possible explanation for this mixed picture is that social competence was operationalized differently in these two quantitative studies. Ullrich-French et al., who reported a growth, operationalized social competence in a way that overlaps with social interaction skills, whereas Laberge et al., who did not report a growth, operationalized social competence in a way that overlaps with social responsibility skills (i.e., respecting others and being polite). The findings from the work by Bruening et al. (2009), who conducted a qualitative case study, were less positive. Although the researchers reported a growth in prosocial involvement in African American and Latina girls who participated in weekly sports sessions and life skills sessions, they only provided a few examples of improvements in behavioral competencies (Bruening et al., 2009). Terry et al. (2014) reported no positive findings on social skills either, as they found that behavioral problems in youth who participated in school-based boxing sessions did not decrease. A final observation was that the studies on the sports programs based on PYD principles reported improvements in social life skills, whereas none of the studies on non-PYD-based sports programs reported or assessed improvements in social life skills.

Conducive conditions
Although many of the included studies referred to program elements as possible explanations for improvements in life skills, only 5 of the 18 studies incorporated research strategies to investigate whether certain program elements were conducive to life skill development. First, on the basis of a quasiexperimental quantitative study, Bonnette et al. (2001) reported that critical-thinking skills increased more for youth participating in a sports program in which the youth themselves had to find solutions for challenges in the sports activities compared with youth participating in a sports program where the coaches prompted these solutions. Second, Anderson-Butcher et al. (2014) found that increased sense of belonging (i.e., feeling comfortable and feeling part of the program) increased the chance for youth developing positive attitudes toward helping other people. Third, according to the parents interviewed by Riley and Anderson-Butcher (2012), the inclusion of a life skill education element, the opportunities for peer interactions, the active and diverse nature of the program, and sports instructors who were caring, personable, and outgoing caused or strengthened life skill development. The teachers and sports coaches interviewed by Armour and Sandford (2013) also reported that positive peer and youth-adult/coach relationships during the sports activities were conducive to life skill development. Another conducive condition that was investigated but did not make a difference for the development of life skills was visits of elite sports role models to the sports program's activities (Armour & Duncombe, 2012).

Discussion
The main aim of this systematic review was to describe the evidence from qualitative and quantitative studies on life skill development in sports programs serving socially vulnerable youth. This review showed that remarkably few studies have been published in this field and that many of the included studies experienced a high risk for bias. Therefore, the results of this review need to be treated with some caution. However, overall, the findings showed that sports programs serving socially vulnerable youth are settings in which socially vulnerable youth can develop life skills, thereby confirming the findings of Lubans et al.'s (2012) review. Each of the 18 included studies reported that at least one life skill improved in the youth who participated in the studied sports program. In contrast to these positive findings, 5 out of 11 quantitative studies reported on life skills that did not improve, and some of the qualitative studies cast doubt on the transfer of improved skills to other settings. This current review expanded on Lubans et al.'s (2012) work by including qualitative studies and more recent quantitative studies. As a result, we found that improvements in cognitive and social life skills were more frequently reported than were improvements in emotional life skills. Finally, this review showed that the studies on life skill development in sports programs serving socially vulnerable youth are very diverse in terms of setting, study design, research method, and reported life skills. However, it seems that the setting in which a sports program is conducted (i.e., school, summer camp, community, and residential care) does not make a difference for whether and which life skills are reported as improving.
The finding that improvements in emotional life skills were reported in fewer studies than were improvements in social and cognitive life skills is contrary to the findings of Eime et al.'s (2013) review. Their review, in which no distinction was made between socially vulnerable and nonvulnerable youth, included many studies that showed a positive relationship between sports participation and emotional life skills. Examples of emotional outcomes that were found to be associated with sports participation but that were not assessed in the studies in this current review are reduced suicidality, reduced mental illness, and increased life satisfaction (Eime et al., 2013). There may be several possible explanations for the low number of studies in this current review that assessed or reported improvements in emotional life skills. First, most of the studied sports programs were based on PYD principles, one of which is to foster cognitive and social competencies (Lerner et al., 2005). Therefore, quantitative research on PYD-based sports programs may more frequently assess cognitive and social skills than emotional skills, and qualitative research may be focused on the program aims when interview data are being coded. For instance, the qualitative studies on sports programs based on the TPSR model (E. Riley & Anderson-Butcher, 2012;Walsh et al., 2010) showed improvements in self-regulation and social responsibility, which are aims of the TPSR model, whereas none of these studies reported improvements in emotional life skills.
An interesting observation was that different life skills were reported as improving in studies using different research methods. Developments in emotional life skills were more frequently reported in quantitative than in qualitative studies, developments in cognitive life skills were equally reported in quantitative and qualitative studies, and developments in social life skills were more frequently reported in qualitative studies. This difference in findings may be caused by the different approaches and research methods, which might steer the researchers' focus toward specific domains of outcomes. Quantitative studies tend to measure the life skills that researchers, policymakers, and/or program staff expect to improve through the sports program, which can lead to a bias whereby specific domains of life skills are omitted. In contrast, in qualitative studies, researchers tend to start with an open mind and attempt not to prompt for specific life skill developments in respondents. However, lack of prompting may lead to bias as well, because youth and their parents participating in qualitative studies may more easily notice and explain developments in social and cognitive skills (e.g., social interaction and self-regulation) than developments in emotional skills (e.g., global self-worth and anxiety).
The second aim of this review was to investigate what is known about conditions conducive for life skill development as identified from the sports programs studied. The conditions found to be conducive in the included studies were a positive youth-coach relationship, sports coaches who encourage youth to deal with challenges that arise in the sports activity, a sense of belonging to the sports program, and the inclusion of a life skills education element. However, as only five of the included studies investigated conditions that may be conducive to life skill development in the sports program, it is not possible to draw firm conclusions regarding the second research aim.
A final major point from this review is that although more and more research is being done in the sport-fordevelopment field (Schulenkorf, Sherry, & Rowe, 2016), relatively few studies have been published that investigate life skill development in sports programs serving socially vulnerable youth. One possible explanation is simply that not much research has been done in this field. Another possible explanation may be the many challenges that researchers in this field have to face, such as high attrition rates, youth workers having priorities other than research, and obtaining parental consent (Whitley, Forneris, & Barker, 2014). Such challenges resulted in increased risk for bias in many of the studies included in this current review. For instance, most quantitative studies lacked a sound comparison group. The comparison group in almost all the quasiexperimental studies included youth who themselves decided not to participate in the sports program. Such a selection bias makes it difficult to compare developments between groups (e.g., Laberge et al., 2012). In addition, when youth themselves decide whether or not to participate in the sports program, it may result in different group sizes. D' Andrea et al. (2013), for instance, compared life skill developments in an experimental group of 62 girls who voluntarily enrolled in the program to life skill developments in a comparison group of 26 girls who decided not to enroll in the program. The qualitative studies included might experience selection bias as well, because youth with negative experiences in the sports program may have dropped out. Consequently, as most interview studies were conducted at the end of a sports program, most of the youth interviewed may have had positive experiences in the sports program.
Given the results of this review, a number of suggestions for further research are offered. First, researchers might consider alternative research approaches to investigate the outcomes of sports programs serving socially vulnerable youth. One alternative might be to adopt a life-course perspective to assess the role of a sports program in youth's lives. This perspective focuses on how the life history of groups or individuals in society may explain differences in well-being. In-depth interviews based on the life-course perspective may encourage adults to explain how they have dealt with the challenges they faced in childhood or adolescence and whether and how sports programs or sports participation helped them to deal with these challenges (Haudenhuyse, Theeboom, & Nols, 2013;Wethington, 2005). Second, researchers might gather longitudinal data from parents, teachers, or program staff members to reduce attrition rates and difficulties around parental consent. An example of such a study was that by Armour and Sandford (2013), who asked teachers to write pupil profiles at several points from the start of a sports program to more than a year after its completion. Third, to increase the comparability of future quantitative studies in this field, researchers may benefit from using general youth development surveys, such as the survey and measurement frameworks developed by Lopez, Yoder, Brisson, Lechuga-Pena, and Jenson (2014) and Vierimaa, Erickson, Côté, and Gilbert (2012). These instruments provide a more holistic picture of youth's development than do questionnaires designed to study one individual outcome. They may also reduce the risk for bias that may result from research tending to measure only the life skills that are expected to improve in youth in a particular sports program.
To study the conditions conducive to life skill development in sports programs, we would recommend that researchers encompass these conducive conditions in their research questions and study designs. In qualitative research, for instance, it could be achieved by asking the interviewees about the elements of the program that they think have led to the life skill development (e.g., Riley & Anderson-Butcher, 2012). In quantitative research, for instance, it could be achieved by comparing the life skill development of youth in a sports program that pays attention to a specific conducive condition with the life skill development in two comparison groups: one group in the same sports program where attention is not paid to this condition and one group not participating in a sports program (e.g., Bonnette et al., 2001). Such quantitative studies provide the opportunity to assess which elements of existing frameworks are conducive to life skill development in sports programs. Examples of elements that might be studied are the implicit versus the explicit approach to the transfer of life skills to other settings (Turnnidge et al., 2014), informal versus organized sports activities (Eime et al., 2013), and collaborative efforts of policymakers, sports organizations, coaches, and parents versus sports programs run by a single organization (Fraser-Thomas et al., 2005).

Limitations
This review is not without limitations. First, although a wide variety of terms related to sports programs, life skills, and socially vulnerable youth were included in the search, the inclusion of additional search terms might possibly have identified more studies. Second, the search terms for life skill development were based on three major life skills (i.e., emotional, cognitive, and social skills), which we chose based on the life skill developments that Turnnidge et al. (2014) linked to sports participation. Other scholars, however, may have categorized life skills in different major categories. Also, whereas we considered responsibility skills to be social skills because they pertain to relationships with other people, others might consider them to be cognitive skills because they overlap slightly with self-control. Third, the search did not include terms pertaining to conditions conducive to life skill development, whereas we looked, as a secondary aim, for conducive conditions within the studies that examined life skill development. Consequently, we may have missed studies that investigated solely conducive conditions for positive sports experiences.

What does this article add?
This review of both quantitative and qualitative studies showed that sports programs are settings in which socially vulnerable youth can develop diverse life skills. Improvements in cognitive and social life skills were more frequently reported than were improvements in emotional life skills. This review also showed that only a few of the included studies investigated the conditions of the studied sports programs that were conducive to life skill development. Finally, we found that still not much research has been published that has investigated life skill development in sports programs serving socially vulnerable youth and that the research that has been published is diverse in terms of setting, research methods, and reported life skills. To provide a better picture of life skill development in sports programs serving socially vulnerable youth, we recommend that researchers in this field consider alternative research approaches, such as adopting a life-course perspective in qualitative studies and using general youth development surveys in quantitative studies.