Extracurricular Activities and Disadvantaged Youth: A Complicated—But Promising—Story

Increased political and research interest in extracurricular activities stems, in part, from the claim that these programs especially benefit disadvantaged youth. However, little literature has synthesized studies across types of disadvantage to assess this claim. This article reviews research on disadvantaged youth in extracurricular programs, including differences by gender, socioeconomic status, race/ethnicity, and immigrant status. Our review reveals a promising, if complicated, picture. Although disadvantaged youth are less likely to participate in extracurricular activities, they often experience greater benefits, depending on the risk status and activity type. Evidence clearly supports expanding access to extracurricular programs for disadvantaged youth.


Introduction
Children from high-risk backgrounds have both the most to gain from afterschool programs in terms of educational opportunity and the least access. (Shumow, 2001, p. 5) OST [Out-of-School Time] programs are an essential component in any strategy to improve the life chances and outcomes for youth, particularly disadvantaged youth. (American Youth Policy Forum, 2006, p. 2) Those students in high-risk situations or struggling in school will tend to benefit the most from quality expanded learning opportunities. (Smink, 2013, p. 56) The advantages of these programs are greatest for children at risk for poor developmental outcomes, such as Latino children of immigrants. (Valladares & Ramos, 2011, p. 1) Discussion about how to best leverage children's development with extracurricular opportunities for learning and growth has taken on new intensity in urban education. As the understanding of urbanicity broadens (Brenner & Schmid, 2015), issues in urban education are relevant not only in major cities but in many communities with urban-like characteristics (Milner, 2012), and to all students who might face oppression (Blanchett, Klingner, & Harry, 2009). These students have multiple influences outside of school that might serve as supports, including extracurricular opportunities (Milner, Murray, Farinde, & Delale-O'Connor, 2015). We use "extracurricular" broadly, to mean activities, whether academic or not, which have not been part of the traditional school curriculum; they are primarily, but not exclusively, outside of school hours and they may or may not be under school auspices or on school property.
One of the most popular ways to package such afterschool opportunities includes community schools. They have become a pillar of school improvement in New York City, which had 215 community schools by the fall of 2017, with plans for more (Zimmerman, 2017). Kentucky made a commitment to community schools early on, and now has over 800 of them, reaching 92% of the state's students (Coalition for Community Schools, 2017). Massachusetts now has 132 such schools and claims they are making considerable headway academically (Farbman, 2015). The National Coalition on Time and Learning is leading a partnership of five states-Colorado, Connecticut, Massachusetts, New York, and Tennessee-which have committed to increasing learning time by 300 hr per year in selected high-poverty schools, allowing for more enrichment activities. Indeed, the once-clear distinction between out-of-school time and in-school learning is becoming blurred. More time in the formal school day often means that activities which were once held afterschool now become embedded in the school day.
Although traditionally liberal foundations like the Ford Foundation and Wallace Foundation have heavily supported the afterschool movement, so have groups like the Broad Foundation, considered a bulwark of conservative causes. Despite some financial exigencies causing programs to retrench, these programs have developed an impressively broad base of support. The largest single funding stream for afterschool is the Department of Education's 21st Century Community Learning Centers initiative, which the Trump administration proposes eliminating altogether for FY 2019 (Peterson, 2018). The growing interest in expanded learning time has several sources, including an increasingly sophisticated body of research attesting to its association with a wide range of academic, social, and psychological benefits (Durlak, Weissberg, & Pachan, 2010;Eccles & Templeton, 2002;Farb & Matjasko, 2012;Knopf et al., 2015;Lauer et al., 2006;Mahoney, Larson, Eccles, & Lord, 2005;Maier, Daniel, & Oakes, 2017;Scott-Little, Hamann, & Jurs, 2002;Vandell, Larson, Mahoney, & Watts, 2015).
As suggested by the opening quotations, part of that interest rests on the belief that enrichment activities offer one way to address the effects of poverty and inequality. Claims of differential effects by race, ethnicity, class, and nationality are of particular importance for urban education. The most disenfranchised urban communities impose layers of social isolation on children that limit their developmental possibilities. It is hoped that extracurricular and extended learning programs can help mitigate those disadvantages (Putnam, 2015).
This article attempts to unpack claims about the potential benefits of afterschool activities for disadvantaged youth. Is there a difference in impact? If so, how does it vary across different types of risk statuses? Gender, poverty, race, and immigration can all pose challenges in urban education-but in different ways. What do we know about patterns and trends in access and participation? What are the implications for policy, practice, and service delivery?

Methods
This article reviews research on extracurricular participation by four categories of risk: gender, socioeconomic status (SES), race/ethnicity, and immigrant status. 1 It was guided by positive youth development (PYD) frameworks, which focus on the positive influences and protective factors that can help at-risk youth not just avoid negative outcomes but to thrive (Lerner et al., 2015). Thus, we conceptualize extracurricular programs as a support for disadvantaged youth that may optimize development, while potentially compensating for or buffering risks.
Studies were identified through Google Scholar, EBSCOhost, PsychARTICLES, and ERIC databases for publications between 1980 and 2017. Search terms included "extracurricular," "out-of-school," "afterschool," and "extended learning"-with terms for risk categories examined. Additional studies were identified from the bibliographies of reviewed articles. Studies were included if they addressed the extracurricular participation of one or more disadvantaged groups. These studies included peer-reviewed journal articles, as well as dissertations and "gray literature" such as white papers and organizational reports, which are common in this research area. Priority was given to more recent studies, to studies that compared disadvantaged and advantaged groups, and to those with an experimental or quasi-experimental design, or methods that accounted for covariates. While random assignment research is becoming more common in this field, it is still rare to have strong grounds for causal inference. Notably, some observers in the field are not convinced that traditional experimental research is the best way to advance this field, noting that programs are often evolving with multiple components that make the "treatment" difficult to distill and randomize (Eccles & Templeton, 2002). This review is organized by category of disadvantage and, when possible, by sports and nonsports activities. Although "nonsports" is a broad category, this distinction was common in the literature, as more research has focused on sports, and comparatively less on "nonsports" programs.

Gender
Students are often channeled and/or select into activities seemingly congruent with their gender 2 -boys into sports and girls into lessons and clubs. However, those activities may not be the most beneficial for them, as research indicates that some organized activities are often more impactful for boys than girls (Posner & Vandell, 1999;Vandell et al., 2015), and this difference varies by activity type. Such trends create an unusual case where girls may be disadvantaged due to less funding for activities they engage in, while boys are disadvantaged for their lower participation in nonsports activities. Major findings on gender are summarized in Table 1 and described below.

Possible Benefits of Sports Participation by Gender
Considerable research suggests sports are associated with both positive and negative outcomes across genders (e.g., Bartko & Eccles, 2003;Darling, 2005;Eccles & Barber, 1999;Fauth, Roth, & Brooks-Gunn, 2007). Although girls participate in athletics less frequently, there are often strong positive academic and developmental corollaries when they do (Farb & Matjasko, 2012;Feldman & Matjasko, 2005). Hanson and Kraus (1998) determined that sports participation was associated with higher math and science achievement among females but not males. Ferris, Oosterhoff, and Metzger (2013) similarly found that among rural youth, sports participation was associated with higher grade point average (GPA) for girls, but not boys.
In noncognitive domains, Fredricks and Eccles (2006) and Simpkins, Ripke, Huston, and Eccles (2005) found that female athletes typically demonstrated higher academic self-concepts than either male athletes or female nonathletes, and female athletes were more likely to have higher achieving friends. Pedersen and Seidman (2004) also determined that higher levels of achievement in team sports predicted higher self-esteem among an ethnically diverse sample of low-income urban adolescent girls. Among a sample of African American, Latino, and White adolescent girls, greater sports participation was significantly related to higher self-worth, sense of body attractiveness, athletic competence, and participation in more extracurricular activities overall (Duncan, Strycker, & Chaumeton, 2015). Randall and Bohnert's (2012) cross-sectional study of urban youth indicated that loneliness levels were lowest among adolescent boys who were involved in sports 3 hr per week, but higher among boys who participated in 7 or more hours of sports weekly. While insignificant, the pattern among girls was opposite, with more time in sports associated with lower levels of loneliness (Randall & Bohnert, 2012). Miller, Sabo, Farrell, Barnes, and Melnick (1999) found that a nationally representative sample of female athletes reported fewer sexual experiences and partners, later onset of first intercourse, higher rates of contraceptive use, and lower rates of past pregnancy than female nonathletes.
Participation has also been associated with greater risk behaviors among males and females, though the picture is complicated. For example, increased aggressive behavior is more likely to be exhibited by student athletes, especially males who play contact sports (Kreager, 2007). Miller, Melnick, Barnes, Farrell, and Sabo (2005) reported that, among eighth through 11th graders, female athletes engaged in more misconduct than female nonathletes over a 2-year period, while male athletes engaged in less misconduct than their nonathlete male counterparts. Male athletes also reported more sexual experiences, more partners, and greater likelihood of using contraceptives than nonathletes (Miller et al., 1999).
With drug and alcohol use, the waters are particularly muddy. In a crosssectional analysis of ethnically diverse, low-income seventh graders, McHale et al. (2005) found that boys who participated in sports were less likely to have used marijuana than male nonathletes, and Fredricks and Eccles (2006) similarly determined that 11th-grade sports participation was related to lower rates of alcohol and marijuana use among boys. However, Crosnoe (2001), Fredricks and Eccles (2006), and Simpkins et al. (2005) found a positive relationship between sports participation among girls and alcohol consumption. Eccles, Barber, Stone, and Hunt (2003) found male and female 12thgrade athletes drank and became intoxicated more frequently than nonathletes, and Hoffmann (2006) found sports participation was associated with increased alcohol use for male and female high schoolers. Mays and Thompson (2009) determined that adolescent male athletes, but not females, were more likely to report heavy drinking than nonathletes.
The identification with an athlete identity may help account for some mixed findings. In a longitudinal analysis, Miller et al. (2005) found that female athletes reported higher grades than female nonathletes, but females who self-identified as "jocks" reported lower grades than female athletes who did not; however, for male athletes, a "jock" identity was not associated with lower GPAs. Miller, Melnick, Barnes, Sabo, and Farrell (2007) reported no direct association between sports participation and problematic behavior, but they did find that students identifying as "jocks" were more likely to report a range of delinquent behavior. Eccles et al. (2003) found that selfidentified "jocks" reported higher rates of drinking as compared with their peers, and Barber, Eccles, and Stone (2001) similarly found that sports participation and espousing the "jock" identity predicted higher rates of alcohol consumption. These findings seem to indicate an important distinction to be made between participating in athletics and developing an athletics-centered identity.

Possible Benefits of Nonsports Participation by Gender
While sports alone has mixed associations, evidence shows positive associations when sports are combined with participation in nonsports activities (Bartko & Eccles, 2003;Duncan et al., 2015;Linver, Roth, & Brooks-Gunn, 2009;Peck, Roeser, Zarrett, & Eccles, 2008;Zarrett et al., 2007). It seems participation in multiple types of programs-or greater activity breadthmay help compensate for negative influences of sports participation, especially for boys.
On average, girls have higher participation rates in arts activities than boys (Schmutz, Stearns, & Glennie, 2016), but studies reported that, for boys only, performing arts participation is linked to lower levels of risk behaviors, including marijuana and alcohol use (Eccles & Barber, 1999;Fredricks & Eccles, 2006). Similarly, Ferris et al. (2013) found participation in church, arts, and music activities was associated with academic achievement for rural boys, but not girls. Kaufman and Gabler (2004) found that participation in yearbook, school newspapers, and/or academic honor societies appeared to increase college attainment for high school boys, but not girls, despite higher participation by girls. DiPrete and Buchmann (2013) suggest that boys who participate in "high-culture" or stereotypically feminine extracurriculars-art, music, dance, and foreign language-experience additional benefits. Drawing on the Early Child Longitudinal Study, DiPrete and Buchmann (2013) found that when boys do participate in high-culture activities, they express attachment to academic values, enjoyment of school, and closeness to teachers at levels closer to girls than to other boys.
In their study of 200 White, working-and middle-class children, McHale, Kim, Whiteman, and Crouter (2004) reported that girls' participation in masculine activities (e.g., hunting, fishing, building, playing with vehicles, or action figures), especially sports, positively predicted math interest. For boys, time in feminine activities (e.g., reading, writing, art, pets, dance, gymnastics, and playing with dolls/stuffed animals), especially music performance, positively predicted their math grades, while participation in masculine activities, especially sports, predicted higher grades in language arts but lower math grades (McHale et al., 2004).

Participation Differences by Gender
Differences in participation patterns by gender seem to be important and persistent (Eccles & Barber, 1999;Feldman & Matjasko, 2005;Holland & Andre, 1987). Girls' activity participation is dispersed across more activity types than boys' (Eccles & Barber, 1999;Feldman & Matjasko, 2007;Fredricks & Eccles, 2006;Jacobs, Vernon, & Eccles, 2005;McHale, Crouter, & Tucker, 2001;Posner & Vandell, 1999), and research consistently demonstrates that boys participate in sports at higher rates and spend more time in unorganized activities than girls (Eccles & Barber, 1999;Feldman & Matjasko, 2007;McHale et al., 2001;McNeal, 1998;Miller et al., 2005;Posner & Vandell, 1999). However, sexual minority males are less likely to report participating in sports than their heterosexual peers (Toomey & Russell, 2013). Although previous research has documented a dramatic increase in girls' sports participation since the passage of Title IX of the Education Amendments of 1972-which protects against sex discrimination in activities that receive federal financial assistance-the gender gap in sports participation increased within all races from the 1980s to the 2000s due largely to the significant increase of sports participation by boys (Shifrer, Pearson, Muller, & Wilkinson, 2015). A gender gap of approximately 25% persisted into the 2000s (Shifrer et al., 2015) and has remained steady or slightly increased since (Child Trends Databank, 2015a).

SES
Although SES is "strongly understudied" in extracurricular research (Farb & Matjasko, 2012, p. 44), studies that do examine SES are consistent on several fronts. Lower SES youth often benefit more from extracurricular participation than those from higher socioeconomic backgrounds. This trend seems true for both academic and nonacademic outcomes, and regardless of whether SES is operationalized by family income, parental education, or composite measures. It is particularly worrisome, then, that lower SES students are less likely to participate in extracurricular activities than their higher SES peers. Major findings on differences by socioeconomic status are summarized in Table 2 and described below.

Possible Benefits by SES
Students from lower SES backgrounds often experience beneficial correlates of extracurricular participation that are equal to or greater than those experienced by their higher SES peers. Several nationally representative and longitudinal studies illustrate these trends. Time spent in academic-focused extracurricular activities in 10th grade was associated with increases in math scores for lower income students, and for students whose parents did not have a college degree, but not for middle-or upper-income youth, or for those with college-educated parents (Morris, 2015). Crosnoe, Smith, and Leventhal (2015) found that extracurricular participation near the transition to high school predicted higher grades in high school, and this trend was stronger for lowincome students. Marsh and colleagues examined the association of 10th-grade extracurricular participation with academic, social, psychological, and risk behavior outcomes; although extracurricular participation was equally associated with most outcomes across SES groups, participation by lower SES youth was more closely associated with higher academic test scores, grades, higher educational aspirations, academic self-concept, social self-concept, greater college enrollment, and greater educational attainment (Marsh, 1992;Marsh & Kleitman, 2002). Lleras (2008) found that participation in sports and academic activities was positively related to educational attainment and earnings for all youth, but the difference in earnings was greater for lower SES youth than others. Dumais (2006) found that low-income elementary school students had the strongest academic benefits from extracurricular participation. Studies with smaller datasets are more mixed. In a sample of 1,047 youth, Fredricks and Eccles (2008) found that the positive correlates of eighth-grade participation were generally equal across a composite measure of SES, with two exceptions: sports participation was associated with more prosocial peers for lower SES youth but not higher SES youth, but greater decreases in depression for higher SES youth than for lower SES youth (Fredricks & Eccles, 2008). However, Fredricks and Eccles (2006) found participation among 11th graders in clubs, arts, and sports was associated with positive outcomes across parental education levels. In a cross-sectional urban sample of 150 ninth and 10th graders, higher dosages of participation were associated with higher levels of loneliness for low-income youth, but not for highincome youth (Randall & Bohnert, 2012).
The school and community contexts in which youth are embedded are closely related to SES and may affect any impact of extracurricular participation. For example, Hull, Kilbourne, Reece, and Husaini (2008) found that participation in nonsports extracurriculars was more protective against emotional distress for Black youth from more disadvantaged neighborhoods than for Black youth in less disadvantaged neighborhoods. Urban, Lewin-Bizan, and Lerner (2009) found highly involved girls from neighborhoods with less resources reported more positive psychosocial development, fewer depression symptoms, and fewer risk behaviors than highly involved girls from better-resourced neighborhoods; however, the reverse was true for boys, such that highly involved boys from better-resourced neighborhoods reported more positive outcomes than highly involved boys from lower resourced neighborhoods. Guest and Schneider (2003) found sports participation was most strongly associated with academic achievement in schools within lowincome communities, while nonsports participation was equally associated with academic achievement for all communities. Hoffmann (2006) found that athletic participation was associated with increased alcohol use over a 2-year period, and this relationship was stronger for females from lower SES schools and for males in higher SES schools, than for other gender-SES groups.

Participation Differences by SES
Unfortunately, lower SES youth are less likely to participate than their higher SES peers. Wimer et al. (2006) looked at two nationally representative datasets and found higher income youth were more likely to participate in all categories of extracurricular activities except tutoring. For example, participation by youth, 12 to 17 years old, was 72% in the highest income quintile, but only 43% for the lowest quintile. Dearing and colleagues (2009) found greater family income was associated with higher participation, and that neighborhood income level partially mediates this relationship between income and participation rates. Pedersen and Seidman (2005) found that lowincome urban youth may have even lower participation rates, potentially due to the additional barriers in these communities. The relationship between SES and participation seems consistent across national (Covay & Carbonaro, 2010;Marsh & Kleitman, 2002;Morris, 2015), regional (Fredricks & Eccles, 2008), and urban datasets (Randall & Bohnert, 2009). 3 Evidence also suggests that these gaps are increasing or, at best, remaining stagnant. Putnam (2015), drawing from four national surveys, reported that the nonsport extracurricular participation gap between highest and lowest income students increased from 1972 to 2002, from a gap of 6.8% to 19.7%. The sports participation gap increased from 19.2% to 24.7% between 1982 and 2002 (Putnam, 2015). 4 Moore, Murphey, Bandy, and Cooper (2014) also used national data to examine the participation of youth ages 12 to 17 above or below 200% of the federal poverty line. From 2003 to 2011, the participation of higher income youth steadily increased, while the participation of low-income youth remained comparatively stagnant, leading to an increase in the participation gap from 19.5% to 28.8% (Moore et al., 2014). Wimer and colleagues (2006) suggested that changes may depend on the activity type, as they found the gap in schoolbased programming actually declined from 1997 to 2002. Nonetheless, these persistent participation gaps are a cause for concern as evidence clearly supports practice and policy that would increase the participation of lower SES youth.

Race and Ethnicity
Extant research on extracurricular participation by racial/ethnic groups is often mixed, with much evidence suggesting that White youth and youth of color experience similar outcomes. When there are differences, White participants often report greater academic outcomes, while non-White participants report greater noncognitive outcomes. Given the potential for positive outcomes, it is concerning that youth of color, particularly Latino/a youth, participate at lower rates. Major findings on differences by race/ethnicity are summarized in Table 3, and described below.

Possible Benefits of Sports Participation by Race/Ethnicity
Studies from nationally representative datasets provide longitudinal evidence of White youth possibly benefiting more from sports participation, especially in educational domains. Lleras (2008) found sports participation was equally associated with higher adult earnings for all racial/ethnic groups, but it was more strongly associated with educational attainment for White and Black youth than for Hispanic and Asian youth. Eitle and Eitle (2002) found that differences may depend on the type of sport. For example, participation in eighth grade football or basketball was associated with lower test scores for both Black and White students, while participation in other sports was associated with higher grades for White students, and lower grades for Black students (Eitle & Eitle, 2002). Latimore, Peguero, Popp, Shekarkhar, and Koo (2018) found that sports participation was not associated with differences in misbehavior by White, Black, Asian, or multiracial students, and only slightly higher reports by Latino athletes. However, sports participation was associated with fewer school disciplinary actions against White and multiracial students, but more school discipline actions against Black, Latino, and Asian students (Latimore et al., 2018). Such findings raise concerns that  Fredricks and Eccles (2006, 2008 Note. SES = socioeconomic status. even if sports participation is not associated with increased misbehavior, schools may respond to White and multiracial athletes more favorably than Black, Latino, and Asian athletes. When studies find positive associations for racial/ethnic minorities, they are often psychosocial or risk behavior outcomes. Eitle, Turner, and Eitle (2003) found greater duration and intensity of sports participation during high school was associated with lower substance use for Black students, but not for White or Hispanic students; sports participation was also associated with higher adult drug use for White students, but not Black or Hispanic students. Stark, Kent, and Finke (1987) found that Black athletes with delinquent behavior in adolescence did not have delinquent behaviors in adulthood, but this change was not true of White athletes. Athletic identity may partially mediate these relationships. Miller and colleagues found that sports participation alone was not associated with delinquent behaviors but identifying as a "jock" was associated with more delinquent behaviors 2 years later for both White and Black students (Miller et al., 2007), and marginally associated with lower grades for Black students (Miller et al., 2005).
In a smaller sample, Eccles (2006, 2008) found that sports participation had mainly equal associations with later outcomes for both White and Black youth across academic, nonacademic, and risk behavior domains; in one exception, sports participation in 11th grade was associated with lower marijuana use for White students but not Black students, and greater political participation 1 year after high school for Black youth, but not White youth. Few recent studies address the intersection of race/ethnicity with gender, though some older studies suggest there may be a need to do so. 5

Possible Benefits of Nonsports Participation by Race/Ethnicity
Evidence also suggests that both White students and youth of color have positive associations with nonsports extracurricular participation. As with sports participation, when differences do appear, White youth often have stronger associations with academic outcomes than youth of color, whereas youth of color have stronger associations with psychosocial outcomes than White youth. Lleras (2008) found that participation in fine arts was more strongly associated with later educational attainment for White youth than for Hispanic or Asian youth, but fine arts were associated with higher earnings for Black and Latino youth and not the other groups. Hull et al. (2008) found that Black and Latino students who participated in nonsports activities had improved psychosocial adjustment 2 years later, whereas White youth who participated did not. Black students in youth programs reported feeling safer and having stronger relationships with staff than White, Asian/Pacific Islander, or Hispanic participants, although feeling safe and relationships with staff were associated with social skill development for all groups except Black youth (Lee, Borden, Serido, & Perkins, 2009). Latimore and colleagues (2018) found that participation in honor societies, school publications, and/or student government was associated with fewer reports of misbehavior by all racial/ethnic groups, and fewer disciplinary actions against White students, but more disciplinary actions against Black, Latino, Asian, and multiracial students. 6 Marsh (1992) examined total number of extracurricular activities, including sports and nonsports, and did not find significant variations between White, Black, and Latino students.
Other studies suggest that greater dosage may yield stronger noncognitive and psychosocial associations for youth of color. With higher dosages of participation, Black youth subsequently spent more time on homework while White youth reported more rigorous course selection, higher educational expectations, and lower risk behaviors (Marsh & Kleitman, 2002). Randall and Bohnert (2009) found that more hours of participation were associated with lower levels of loneliness for Black youth, while longer duration in activities was associated with lower levels of victimization for Latino students, but with higher victimization for Asian students. Brown and Evans (2002) found that hours of participation were associated with greater school connection, and that this relationship was strongest for Hispanic students, followed by Black and then White students. Mahoney, Harris, and Eccles (2006) found that White youth reported higher reading achievement and lower smoking rates with the most hours of participation, but that Black youth reported higher emotional well-being and self-esteem with the most hours of participation. Greater duration and intensity of sports participation were associated with lower substance use for Black youth, but not White or Hispanic youth (Eitle et al., 2003). 7

Participation Differences by Race/Ethnicity
Although a decade old, most data consistently suggest that Latino youth participate at the lowest rates, followed by Black and then White youth. Wimer et al. (2006) found, of adolescents aged 12 to 17, 65% of White youth participated in any organized activity, followed by 58% of Black youth, and 43% of Latino youth. This trend is consistent across activity types, except that Black students report higher participation rates in school-based programs and tutoring programs (Marsh & Kleitman, 2002;Wimer et al., 2006). White and Black youth participate in extracurriculars around 5 hr per week (Mahoney et al., 2006) and have a similar breadth of participation (Bartko & Eccles, 2003;Feldman & Matjasko, 2007). Similar participation trends are also seen in other nationally representative datasets (e.g., Feldman & Matjasko, 2007;Hull et al., 2008;Kann et al., 2014), though more recent studies are needed. 8 In summary, studies generally find equal and positive outcomes from participation for both White youth and youth of color. When there are exceptions, there are often stronger academic outcomes for White youth, and stronger psychosocial and nonacademic outcomes for youth of color. Higher dosages of participation may be particularly advantageous to Black and Latino/a youth for social integration, school connectedness, and general psychosocial adjustment. Despite consensus that extracurricular programs can be important supports for Black and Latino youth (e.g., Fashola, 2003;Fredricks & Simpkins, 2012;Woodland, 2016), more research is needed on how programs can best support youth of color. For example, we would caution that there seems to be little research on programs that deal directly with issues of race, ethnicity, and culture. Ethnic and multicultural studies in the school curriculum are associated with positive outcomes on a wide array of important social and academic indicators, including a sense of agency, engagement, higher order thinking, high school graduation, grades, motivation, and writing skills (Cabrera, Milem, Jaquette & Marx, 2014;Dee & Penner, 2016;Fashola, 2003;Sleeter, 2011). There are afterschool programs that are similar-the Children's Defense Fund, Freedom School, or organizations like New York's Brotherhood Sister Sol-but we have little systematic knowledge about their effects.

Immigrant Status
Foreign-born Americans and their children will be a major source of population growth in the United States over the next 50 years (Pew Research Center, 2015). Children who have at least one parent born outside the United States presently make up almost one quarter of the children in the country (Capps, 2001;Valladares & Ramos, 2011) and an increasing share of the school population (Cherng, Turney, & Kao, 2011;O'Hare, 2004). Few studies specifically examine immigrant extracurricular participation, and those that do primarily focus on Latino immigrants and nonrefugee populations. 9 Overall, compared with children in native-born families, children in immigrant families are less likely to participate in afterschool activities (Cherng et al., 2011;Greenberg, 2014;Peguero, 2010;Reardon-Anderson, Capps, & Fix, 2002). However, this may depend on the immigrant group and activity (Simpkins, O'Donnell, Delgado, & Becnel, 2011). For example, Yu, Newport-Berra, and Liu (2015) found that, compared with White nonimmigrant youth, White immigrant youth were more likely to participate in clubs and community service, Black immigrant youth were more likely to  (2007) Alcohol Use: Fredricks and Eccles (2006) Academic and Noncognitive: Marsh (1992) Academic, Social and Risk Behaviors: Eccles (2008, 2010) When Black or Latino youth benefit more from participation than White youth, it is in terms of nonacademic outcomes.
Delinquency: Stark, Kent, and Finke (1987) Income: Lleras (2008) Noncognitive: Hull et al. (2008) Greater dosage is associated with greater benefits for Black and Latino youth than for White youth.

Brown and Evans (2002)
Academic & Risk Behaviors: Mahoney, Harris, and Eccles (2006) Marsh and Kleitman (2002) Academic: Cooper, Valentine, Nye, and Lindsay (1999) participate in sports, and other non-Hispanic immigrant youth were more likely to engage in community service. Family factors such as parental income, structure, supervision, and communication influence immigrant youth's extracurricular participation, and this effect is stronger for Latino immigrants (Jiang & Peguero, 2017). Barriers to participation include family responsibilities, limited financial resources, language, unfamiliarity with the American school system, lack of transportation, safety concerns, the negative influences of peers, and fear of racism or discrimination (Bejarano, 2007;Borden et al., 2006;Lin et al., 2016;Simpkins, Delgado, Price, Quach, & Starbuck, 2013;Simpkins et al., 2005;Vernez, Abrahamse, & Quigley, 1996). Immigrant youth also may be more likely to participate when they are enrolled in racially diverse schools (Okamoto, Herda, & Hartzog, 2013). As most literature does not compare outcomes of immigrant with nonimmigrant youth, the remainder of this section focuses on the possible benefits of these programs within immigrant youth samples. Major findings on participation by immigrant youth are summarized in Table 4, and described below.

Possible Benefits of Participation by Immigrant Youth
Research suggests participation by immigrant youth is associated with positive academic and psychosocial outcomes. In a survey of 468 Latino 11th graders, Camacho and Fuligni (2015) found that first-generation immigrant youth were less likely to participate in academic activities than their third-generation peers; however, participation was associated with higher academic achievement and engagement, and gains in GPA following participation were greater for first-generation immigrants than thirdgeneration participants. For rural immigrant Latino children, Riggs (2006) found that higher dosage was related to increased social skills and decreased behavior problems. Some studies suggest that civic participation and social justice programs may be one area where immigrant youth, particularly Latinos, are increasingly involved (Perez, Espinoza, Ramos, Coronado, & Coates, 2010;Suárez-Orozco, Hernández, & Casanova, 2015;Yu et al., 2015). Qualitative data emphasize the importance of supportive adults for Latino immigrant youth. Gonzales (2011) found that Latino youth were often reluctant or unable to participate in extracurricular clubs and frequently needed the intervention of a teacher to get involved. Diversi and Mecham (2005) also found that Latino immigrant youth benefited from relationships with supportive adults which fostered academic engagement, cross-cultural relationships, higher grades, and stronger connection to school. Cooper, Valentine, Nye, and Lindsay (1999) likewise concluded that adult mentors in extracurricular programs served an important integration role for Latino immigrants, helping youth cultivate bicultural and language skills. Camras (2004) found relationships with adults helped build the social affiliation and social capital of immigrant youth, and Ettekal, Gaskin, Lin, and Simpkins (2016) found the participation and engagement of Latino immigrant youth depended on adult staff establishing and maintaining a culture of respect.
Although these studies suggest that afterschool programs can be helpful to Latino immigrant youth, others suggest limited benefits or even added risk. Park, Lin, Liu, and Tabb (2015) found that extracurricular participation was not associated with any differences in grades, school work, behavioral problems, or suspensions for Latino children from immigrant families. The authors argue that existing afterschool programs are not a strong fit for Latino youth from immigrant families, and programs need to incorporate culturally appropriate components (Park et al., 2015). Other data suggest that first-generation youth who participate in extracurricular activities are more likely to experience violence (Jiang & Peterson, 2012).

Conclusion
For the risk statuses examined in this review, it seems that extracurricular participation is frequently associated with different and sometimes stronger outcomes. We wish to stress the points listed in Table 5.
We clearly need more research in this field generally, and on the participation of non-Latino immigrant youth, sexual minority youth, youth in foster care, and youth facing other social risks (see Note 1). More research should address intersectionality and issues of causal pathways and should be embedded in theories of youth development. On the practice side, it seems clear that schools and youth programs should be monitoring the overall participation of  Peterson (2011) Park, Lin, Liu, andTabb (2015) Participation is associated with improved academic outcomes for immigrant youth. Camacho and Fuligni (2015) Cooper, Valentine, Nye, and Lindsay (1999) Espino et al. (2004) Fabiano, Pearson, Reisner, and Williams (2006) Huang, Kim, Marshall, and Pérez (2005) Valdés (1998) Vandell, Reisner, and Pierce (2007. Welsh, Russell, Williams, Reisner, and White (2002) Park et al.
Participation is associated with English language learning.  Nonsports extracurricular activities may have academic and nonacademic benefits for both genders. Sports are more of a mixed bag, as they are associated with positive outcomes for both genders, but also possibly stimulate risk behaviors, at least for boys. Risk outcomes may be mediated by whether youth develop an identity as "jocks." Boys' activity breadth narrows as they age, and they generally engage in more unstructured activities than girls. Heterosexual girls and sexual minority males participate in sports at lower rates than heterosexual males, and the gender gap appears to have increased over time, despite efforts like Title IX. Youth prefer activities that are congruent with their gender, but there may be greater benefits when they try activities that are gender atypical. Both researchers and practitioners should think about whether this is one case of a more general pattern in which there are particular benefits for youth who cross social boundaries-White students in the gospel choir, low-income students on the golf team, male students on the yearbook staff. This may be particularly important for boys, whose concentration in team sports may not be optimal. It seems very likely that the possible benefits for poor children are greater than those for the more well-to-do, and this extends across a range of outcomes: educational, psychological, social, and behavioral. Lower income youth participate at lower rates, however, and evidence suggests that the participation gap has been growing over several decades, though this may vary by activity type. Although more mixed, findings on racial/ethnic differences suggest that youth of color might benefit as much as or more than White youth in terms of noncognitive outcomes, and higher dosages of participation may be especially advantageous for youth of color in terms of both academic and nonacademic outcomes. Few recent studies address the intersection of race and gender. 5 While we know surprisingly little about the activities of Black girls and Latinas, we do know they are less physically active than other girls (and boys). Latino/as, and especially Latino boys, continue to have significantly lower rates of participation. Some of the groups that might benefit most, participate the least. Limited work, much of it focused on Latino/as, suggests extracurricular participation can support the integration of immigrant youth, influencing language development, academic engagement, social competence, and behavioral issues. The work suggesting that schools may discipline non-White participants more than nonparticipants implies that, whatever benefits may come from participation, participation might not change the way institutions react to stigmatized youth. Indeed, it may put them at greater risk, perhaps by making them more visible.
particular subgroups-including Latino/as, immigrants, girls, and lowincome youth-and the particular types of activities in which they participate. If we want to put young people in the most advantageous position, it may be necessary to do a great deal more targeted recruiting, steering youth toward activities they might never consider on their own. While there is still a need to make more opportunities available to disadvantaged youth, we have to be proactive about making sure they take advantage of the opportunities that do exist. Although beyond the scope of this review, scholars and practitioners have begun identifying indicators of quality and best practices in this field (Holstead, Hightower King, & Miller, 2015;Little, Wimer, & Weiss, 2008;National Research Council and Institute of Medicine, 2002). It would be important to know whether the quality of program implementation varies by the kinds of disadvantaged statuses examined here. The idea that improving the academic performance of disadvantaged children is the best way to put them on a path to better adulthoods is rarely examined critically in the social sciences, or in urban education specifically. Some recent research, however, argues that the life outcomes of young people are more strongly driven by "character skills" than by academic skills (Heckman, Humphries, & Kautz, 2014), the kinds of skills presumably more likely to be developed in extracurricular activities. There is also evidence that while investments in cognitive skills pay weaker dividends with older students, noncognitive skills can be effectively developed with that population (Cunha, Heckman, & Schennach, 2010). Putnam (2015) contends that one of the disadvantages children from less affluent homes face is that they have fewer informal mentors, despite being more likely to express the desire to have mentors; afterschool programs may provide necessary opportunities for mentoring. It is cause for concern, then, that we have evidence of dramatic and growing gaps in private enrichment expenditures for high-and low-income children-US$8,800 per year per capita for high-income children against US$1,300 a year for low-income children (Kaushal, Magnuson, & Waldfogel, 2011). Debates about neoliberal educational policy have rarely focused on extracurricular activities. This review suggests we should be concerned that the privatizing of developmental opportunities, through austerity budgets and pay-to-play policies, will further exacerbate participation gaps, depriving disadvantaged youth of possible pathways to a more satisfying and productive adulthood. The evidence here comes down clearly on the side of deeper investments in the kinds of experiences that will give more children-especially disadvantaged youth-a chance to find and develop their gifts.

Author's note
Ryan D. Heath is now affiliated to Syracuse University, Syracuse, IL, USA.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by the Spencer Foundation and the Sebring-Lewis Foundation.

Notes
1. Although not reviewed in this article, our search found studies on the participation of youth with other forms of disadvantage. These risks included foster care (West-Bey, 2014), sexual minority status (Toomey & Russell, 2013), single parent families (Brown, 2011), academic ability (Marsh, 1992;Marsh & Kleitman, 2003), lower psychosocial adjustment or social skills (Mahoney, Cairns, & Farmer, 2003;Peck, Roeser, Zarrett, & Eccles, 2008), delinquency (Taheri & Welsh, 2016), and latent classes of risk (Mahoney, 2000;Mahoney & Cairns, 1997). 2. Although we recognize gender as a multidimensional construct, this review found no extensive literature on gender identities other than male and female. 3. One exception to these trends is found in a report by the Afterschool Alliance (2009), which reports that students qualifying for free or reduced lunch participate in afterschool programs at 18%, as opposed to 14% for students not qualifying for free or reduced lunch. This finding is contradictory to other reports, as were the Afterschool Alliance's (2009) findings on racial/ethnic minority participation rates. 4. Although drawn from nationally representative surveys, these studies only include participation in school-based extracurriculars by non-Latino White students (Putnam, 2015). 5. Several older studies consider the intersection of gender with race/ethnicity and suggest that White males may have stronger academic outcomes from sports, while Black female athletes may have increased academic risks from sports. Spreitzer (1994) found the relationship between sports and educational attainment was strongest for White male athletes, moderate for White females, and weakest for Black and Hispanic youth. Sabo, Melnick, and Vanfossen (1993) found that sports participation was associated with higher educational attainment most strongly for White males, moderately for suburban females and rural Hispanic females, but not related to educational attainment for Black students or Hispanic males (Sabo et al., 1993). Tracy and Erkut (2002) found sports participation had direct associations with higher self-esteem for White males and Black females, but not for White females or Black males. Other studies found sports participation was negatively associated with attitudes and achievement in science and mathematics (Hanson & Kraus, 1998) and occupational aspirations and attainment (Sabo et al., 1993) for Black female athletes. Although dated and less rigorous, these studies suggest a need to investigate intersectionality. Only one study reviewed had more current data, and it found that greater sports participation was related to less depression for Latino and White girls, but not for African American girls (Duncan, Strycker, & Chaumeton, 2015). 6. In contrast to studies finding differential associations by race/ethnicity, Fredricks and Eccles (2006, 2008 studied the association of extracurricular participation with a host of cross-sectional and longitudinal outcomes, including grade point average (GPA), self-esteem, resilience, depression, prosocial peers, and risk behaviors. Within this sample of 1,074 youth, participation in school clubs and service/religious activities was equally advantageous to both Black and White students for most outcomes, with little evidence of differences crosssectionally, 4 years later, or 6 years later. 7. One disconfirming study on dosage by Fredricks and Eccles (2006) found a greater breadth of participation was associated with lower substance use for White youth, but slightly higher use among Black youth. A second study by Cooper, Valentine, Nye, and Lindsay (1999) found that time spent in extracurricular activities was associated with higher test scores equally for both White and Black students. 8. Appendices in reports by Child Trends Databank (2015a, 2015b, 2015c) report raw participation rates of White, Black, and Latino youth from 1995 to 2011, but do not formally analyze or discuss these trends. While the sports participation gap between Black and White youth closed between the 1990s and 2000s, it since may have widened; the sports participation of Latino youth is only reported from 2009 to 2011, and it appears to be increasing in that time (Child Trends Databank, 2015a). In performance (Child Trends Databank, 2015b) and volunteering activities (Child Trends Databank, 2015c), the participation trends appear similar for White, Black, and Latino youth. 9. We note that refugee youth face specific challenges that differ from nonrefugee immigrant youth, and few studies address refugee extracurricular participation (e.g., Mendenhall & Bartlett, 2018).