The class “extracurricular activities” (EA) constitutes a positive youth developmental asset (Durlak et al., 2010; Eccles and Gootman, 2002; Eisman et al., 2016; Farb and Matjasko, 2012; Mueller et al., 2011) and covers a broad range of categories that share some essential elements (Hansen et al., 2010). Indeed, competent adults supervise these activities, which often involve peer interaction. They also have regular attendance schedules, offer practical learning opportunities and enable young people to spend time engaged in their own interests (Hansen and Larson, 2007).

EA does not form part of the school curriculum and, unlike formal education, student participation is voluntary. Adolescents often develop meaningful relationships with their peers and their instructors. This fact creates an appropriate context in which to develop identity, initiative, and social skills (Hirsch et al., 2011; Larson et al., 2006). These aspects help in theorizing about organized activities and contribute to different processes of adolescent development.

Moreover, investigations of the types of activities that involve youth in their free time have identify the relevant components that promote positive development in supervised/structured environments or, specifically, in extracurricular programs or activities. Positive youth development programs incorporate manifold elements that promote such development, with importance directed toward adult–adolescent relationships that tend to involve young people over time (Roth et al., 1998).

Type of activity

Studies of different types of EA address various development experiences (Eccles and Gootman, 2002), such as sports, performing arts, service clubs, and faith-based youth groups (Vandell et al., 2015). Research shows differences between sports and artistic EA, based on the results they promote in youth (Hansen et al., 2003; Im et al., 2016). Thus, evidence shows that artistic activities improve adolescent adjustment, as well as the participants’ self-knowledge, self-discipline, and artistic talents (Hansen et al., 2003).

Sports EA not only prevent risk behaviors, but also improve the social and academic abilities of youth (Darling, 2005). However, collective sports contribute to a lower level of academic, social, or preventive skills promotion than other types of EA (Hansen et al., 2010; Wilson et al., 2010).

Bartko and Eccles (2003) classify adolescents very involved in sports, spending more time with friends, and performing other EA in different groups, with high participation rates in school clubs, tasks, and reading for pleasure. For Im et al. (2016), participation in sports and artistic activities predicts an increase of self-efficacy in academic competence, so EA provides a context in which students can face and overcome challenges and increase the level of their skill, thus building trust.

Based on the approach of Mahoney et al. (2005), Reading books for pleasure would not be an organized extracurricular activity. In fact,

“these activities are generally voluntary, have regular and scheduled meetings, maintain developmentally based expectations and rules for participants in the activity setting (and sometimes beyond it), involve several participants, offer supervision and guidance from adults, and are organized around developing particular skills and achieving goals” (Mahoney et al., 2005, p. 4).

Nonetheless, Reading books it is very relevant for youth development, relating to greater participation in a wider breadth of activities, as well as greater engagement in both extracurricular and curricular tasks (Bartko and Eccles, 2003).

Activity engagement

For EA to function as more effective assets requires youth engagement (Weiss et al., 2005) though factors such as breadth, intensity and duration (Bohnert et al., 2010; Busseri and Rose-Krasnor, 2009; Busseri et al., 2006). The benefits of participation in EA, both in the curricular field and in the personal development of the young person, depend on the characteristics of their participation experience. In turn, such experiences depend on their engagement, whether psychological (e.g. duration) or behavioral (e.g. breadth, intensity). However, a consensus on the conceptualization of participation should specify the aspects of behavioral and psychological engagement (Eisman et al., 2016).

Breadth of activities

Breadth refers to the number of EA that involve youth. The study of Breadth provides a broad description of the range of student skills and interests (Bohnert et al., 2010; Busseri and Rose-Krasnor, 2009; Busseri et al., 2006; Linver et al., 2009).

The range of activities refers to the participation in different activities in a certain period and may refer to the total number of activities in which a young person engages (Busseri and Rose-Krasnor, 2009) or the total number of different types of activities (Eccles and Barber, 1999). Contrary to participation in a single activity, a wide range of participation allows young people greater diversity of learning experiences, supportive interactions with adults, and broad networks of peers (Bohnert et al., 2010; Eccles and Barber, 1999; Vandell et al., 2015).

Duration of activities

The duration or consistency of participation in EA refers to the maintenance over time. “Dosage has been conceptualized in terms of consistency and continuity over time, measured as the proportion of periods in which youth engage in activities” (Pierce et al., 2013; in Vandell et al., 2015).

Longitudinal studies have measured consistency of participation among both elementary and middle-school children (see National Institute of Child Health and Human Development Early Child Care Research Network, 2004; Pierce et al., 2013) and in adolescents (see Mahoney et al., 2003; Zaff et al., 2003). The participation in EA over the years could increase the benefits for youth development, as a result of greater exposure to such opportunities (Eisman et al., 2016; Tudge et al., 2009).

Differences in adolescents’ activity participation

Sex differences in adolescents’ activity participation

Sex relates to participation in EA, with girls showing more behavioral engagement in EA except in sports (Denault and Poulin, 2009b; Eccles et al., 2003; Kleiner et al., 2004). Indeed, boys are more likely than girls to participate in sports, while girls are more likely than boys to participate in classes and clubs (Kleiner et al., 2004) or artistic EA (Luthar et al., 2006). The impact of EA on positive development is higher for boys (Pierce et al., 2010; Urban et al., 2009). No evidence indicates differences in participation according to sex, rather than age (Eisman et al., 2016). In any case, evidence of gender-related differences in the effects of EA is scarce.

Regarding risk behaviors, according to Eccles et al. (2003), participation in a competition sports team in late adolescence for both sexes is associated with higher alcohol consumption, the opposite of participation in artistic activities.

Stage differences in adolescents’ activity participation

Eisman et al. (2016) argue that the stage of development of the young person determines the importance of each type of engagement. In early adolescence, people tend to participate in a wider range of EA. Indeed, the breadth of participation could be more relevant in early adolescence than in late adolescence, because exploring different interests and strengthening bonds with peers characterizes early adolescence (Bohnert et al., 2010; Busseri and Rose-Krasnor, 2009). Also, as mentioned, the breadth of participation enables more learning opportunities and broader networks with different groups of peers and supportive adults (Vandell et al., 2015).

The young persons’ persistence in an activity depends on different factors, such as the activity itself, personal characteristics, and interests, and other environmental aspects, such as family circumstances, peers, and school or community characteristics (Vandell et al., 2015). Accordingly, as progress is made in developmental stages, adolescents develop greater control over their use of time (Fredricks and Eccles, 2010) and refine their personal interests. Consequently, in middle-to-late adolescence, young people participate in fewer activities, but with more intensity (Busseri et al., 2006; Denault and Poulin, 2009a) and interest.

Indeed, leaving a type of activity may reflect the progression of development. Thereby, young people begin testing different EA, then focus on a smaller number of them as their interests become more defined (Badura et al., 2016; Rose-Krasnor et al., 2006), ranging from participating in a wide range of activities to a small number over time (Denault and Poulin, 2009a; Eccles and Barber, 1999; Rose-Krasnor et al., 2006). In fact, participation in EA cannot change throughout adolescence (Eisman et al., 2016; Zaff et al., 2003) or change constantly along this stage (Farb and Matjasko, 2012), despite disagreements about whether such participation increases (Mahoney et al., 2003) or decreases (Denault and Poulin, 2009b).

Parental education level differences in adolescents’ activity participation

Taking account of the context in which young people live is essential to understand the characteristics of their participation in EA. Research on differences in participation have been focus more in terms of socioeconomic factors (Luthar et al., 2006; Pedersen, 2005; Quinn, 1999;Vandell et al., 2015) than in differences related to parental education (Kingdon et al., 2017).

Parental education level is associated with participation in EA, in that youth whose parents have a higher educational level tend to participate more in activities than those whose parents have lower educational levels (Bartko and Eccles, 2003; Eisman et al., 2016; Feldman and Matjasko, 2007; Linver et al., 2009; Vandell et al., 2015), and parental education level also predicts their children’s duration in activities (Eisman et al., 2016).

EA associations with academic achievement

The relationship between participation in EA and improvements in academic achievement have been extensively studied. How adolescents decide to manage their free time is a protective factor related to academic achievement in higher grades, as well as to recovering from low GPAs (Eccles et al., 2003; Linver et al., 2009; Peck et al., 2008; Roth et al., 2010).

Participation in EA also relates to furthering the adolescent’s permanence in the educational system by improving their behavior and school attendance (Simpkins et al., 2004). Participation in EA improves other relevant aspects of curricular success, such as lower school dropout rates (Mahoney, 2000) and the school climate, in terms of associations with friendship and prosocial behavior with peers, as well as less aggressive behavior toward them (Eccles and Templeton, 2002).

Participants in sports activities experience lower academic achievement than those who participate in artistic activities (Eccles et al., 2003). Reading books is a protective factor in preventing school failure, and it improves academic achievement in early (Kingdon et al., 2017) and late adolescence (e.g. Horbec, 2012; Whitten et al., 2019). Adolescents reading books during their extracurricular schedule has been related to higher academic aspirations (McGaha and Fitzpatrick, 2010). Conversely, lack of reading has been related to difficulties in the transition to college (Rubin, 2008).

Most of the research that associating EA with academic achievement shows the latter as an outcome that participation improves (Eisman et al., 2016; Linver et al., 2009). Although both breadth and duration have been related to academic achievement (Palma et al., 2014), behavioral engagement (i.e. breadth) shows more predictive ability (Barber et al., 2005; Eisman et al., 2016; Fredricks and Eccles, 2006). However, duration—evaluated as consistent participation in a wide range of EA—has also predicted higher grades and academic achievement in studies that control for sex and parental education (Darling et al., 2005; Zaff et al., 2003).

The traditional controversy over whether breadth contributes to improving academic achievement (Linver et al., 2009) has given way to recent findings that indicate the relation of breadth to improvement in early adolescence (Eisman et al., 2016; Roth et al., 2010). However, the same would not happen in late adolescence, when duration is the more consistent predictor (Eisman et al., 2016).

Regarding the relationship between participation in EA and academic achievement in terms of sex and age, younger girls show the highest academic achievement (Kingdon et al., 2017; OECD, 2015), depending on the type of activity. For instance, an association has been found between reading books and academic achievement in early adolescence (i.e., schoolchildren between 12 and 14 years old) (Kingdon et al., 2017).

Aims and hypothesis

Despite the abundant literature on the effects of different EA on academic performance, empirical evidence is lacking for the relationship between extracurricular variables, such as breadth and duration, and academic achievement in Hispanic (including Spanish) contexts. Therefore, we aimed to investigate in such a context two groups of noncurricular factors—EA and sociodemographic variables—and a key curricular variable, academic achievement. In particular, we consider sex, age (distinguishing between early middle and late adolescence) and parental education level as sociodemographic variables.

On the other hand, as variables of EA, we take into account the breadth—understood as Busseri and Rose-Krasnor (2009) state it—and duration of participation, the type of activity (i.e., individual sport, collective sport and arts), and add an informal activity traditionally related to academic achievement, the reading of books. Associations of EA and sociodemographic variables with academic achievement were analized.

The hypotheses of the investigation are:

  1. 1.

    Types of activities (i.e., reading of books, artistic, and individual or collective sports) are associated with academic achievement, controlling for sex, age, and parental education level.

  2. 2.

    Duration of EA is associated with academic achievement, controlling for sex, age, and parental education level.

  3. 3.

    Breadth of EA is associated with academic achievement, controlling for sex, age, and parental education level.



Participants were recruited at 10 schools randomly selected from among all secondary schools in the province of Zaragoza, in Spain. We requested participation from students in grades 7, 9, and 11 (around 12, 14, and 16 years old, respectively). In total, 1148 students completed the survey, with balanced distributions by sex (50.2% female) and age (distribution of ages can be found in Table 1).

Table 1 Characteristics of the sample.


EA (Hermoso et al., 2010) contains descriptive data about performing organized activities after school hours. From among adolescents who participated in such organized activities, data for a series of dummy variables were collected, covering the number of courses underway, the type of activity (sports and/or arts), and other items related to students’ perceptions of participation. Experts assessed and critiqued the questionnaire, and it fulfilled the requirements of external, internal, and content validity.

Sociodemographic variables data on sex, age, and parental education level were collected.

Academic achievement attends to a self-report of the previous course GPA.


This study was carried out in accordance with the recommendations of the Council of the British Educational Research Association in its second edition of the Ethical Guidelines for Educational Research (2011). Subjects received no compensation for participating in the study. Compliance with these ethical standards was guaranteed at all time.

The objectives and characteristics of the study were explained to the principals and counselors, who agreed to participate. Afterwards, they transferred the study objectives and questionnaires to the tutors from the different groups. Prior to completion, families were informed by a letter of the purpose of the study and procedure, and participants’ anonymity was ensured. In the same letter, the volunteers were informed of participation and the possibility of excluding from the activity those children whose families did not consent to participation, given that the data was collected during class time.

Data analyses

The age of the participants was dichotomized at the median, which resulted in one group between 12 and 14 years of age and the other between 15 and 18 years of age. For the evaluation of parental education level, three categories were considered: primary basic studies, middle studies, and university studies. The participant was assigned to one according to the highest level either parent had reached. For sports activities, participants were asked if the child was currently involved in individual or collective sports activities.

For the identification of artistic activities, if they were currently participating in dance, theater, music, or plastic-arts activities, any positive response on any of these activities was considered an affirmative response for this type of activity in its entirety. Also, they were asked directly if they were currently dedicating their time to reading books. The time devoted to EA was determined by asking the respondents to report during how many courses they had carried out these activities.

Finally, regarding the breadth of EA, the definition of Busseri and Rose-Krasnor (2009) was adopted, which proposes breadth as the total number of activities in which a young person engages. The response variable related to academic achievement was obtained by asking for the final grade in the immediately preceding course, reported on a scale from 1 to 8. However, for the descriptive analysis, this variable was categorized as “deficient” (below minimum approval level), “sufficient” (between minimum approval level and up to 5 units), or outstanding (higher than 5 units).

For the evaluation of the proposed hypotheses, a multilevel model of main effects was evaluated, where covariates and evaluation factors were taken as fixed effects, and the school to which the student belonged was categorized as a random effect. The estimation method was through restricted maximum likelihood with the Satterthwaite method for calculating the degrees of freedom. The residuals of the model were evaluated for their normality by means of skewness and kurtosis criteria. A significance value of p < 0.05 was used. For post hoc comparisons, t-tests protected by Fisher were used in cases where the factor had more than two levels; for the rest of the cases, the F-test was considered sufficient. Due to the differential effect of the age of the respondents on the confounding and response variables, the proposed model was estimated independently for each age group.


In total, 549 (47.8%) males and 599 (52.2%) females were recruited, a greater proportion of whom were 12–14 years of age, with parents who had a high educational level (see Table 1). Regarding the response variable “academic achievement”, the three categories presented show a similar distribution; however, girls show better achievement than boys by more than 10 percentage points in the superior category.

The analysis of variance (Anova) of the multilevel model, proposed for the evaluation of academic achievement as a dependent variable, shows that the responses related to the evaluated variables corroborate the differentiated behavior by age group. For the younger group (12–14 years old), sex, reading books, and duration of EA are significant factors associated with academic achievement, contrary to the results for the older group (15–18 years old) for whom collective sports activities and breadth of EA was significantly related to academic achievement. Only parental education level presented a similar effect for both age groups (Table 2).

Table 2 Anova of the multilevel model for extracurricular activities (EA) and academic achievement as the dependent variable.

The analysis of means showed that for the sex-related factor, between 12 and 14 years of age, the girls had better academic achievement (see Table 3 and Fig. 1a). For the level of education of the parents, those young people whose parents only reported levels of primary education showed the lowest academic achievement, compared to those with both medium levels and high levels of parental education. Regarding collective sports activities, the practice of these activities by the older students affected positively to academic achievement, as Fig. 1b shows. On the contrary, the effect of reading books was found among the younger ones as a promoting factor of academic achievement, also shown in Fig. 1b. Finally, for the variables of time and breadth, a differential in behavior was found again. For the younger students, the time spent most affected their academic achievement, by increasing it. For the older students, breadth was a predictor of academic achievement—in this case, the low level of dispersion is the weaker association, compared to the medium and high levels that present the best values (see Fig. 1c).

Fig. 1: Means of academic achievement grouped by age sets, predictors, and confusing factors.
figure 1

a Sex and parental education level; b breadth and duration of EA; and c collective sports activity and reading of books.

Table 3 Comparison of means for levels of factors evaluated of the multilevel model for each age group with extracurricular activities (EA) as the dependent variable.


The main contribution of this paper is the study of these characteristics in Hispanic, concretely, Spanish contexts, which is not very common in the scientific literature. Our results indicate that adolescents’ choices of EA participation are associated with academic achievement, as other investigations have shown (Badura et al., 2016; Bartko and Eccles, 2003; Eisman et al., 2016; Linver et al., 2009), but this association differs according to sex, age, type of activity, and parental education level. As the previous literature shows, boys participate more in sports activities while girls do more in artistic activities (Darling, 2005; Kleiner et al., 2004; Luthar et al., 2006), and participation is higher in early adolescence than in late adolescence (Denault and Poulin, 2009b, 2009a; Rose-Krasnor et al., 2006).

Regarding the control variables in the prediction of academic achievement, the youngest girls have the highest grades, as has been found previously (Badura et al., 2016; OECD, 2015). On the other hand, a higher level of parental education is associated with higher academic achievement, in both early and late adolescence (Bartko and Eccles, 2003). The study by Roth et al. (2010) produces the same result, as well as an association with a smaller number of behavioral problems, but only when youths with high levels of participation were compared with peers who did not attend EA. Kingdon et al. (2017) found that reading and involvement of mothers are protective factors against the decrease in academic achievement in boys, as well as a predictor of future achievements, especially among low-income groups.

Addressing the first hypothesis (i.e., type of activity is associated with academic achievement), academic achievement relates negatively with collective sport EA in late adolescence. This is in line with the previous literature where sports activities relate to lower academic achievement than other types of EA, such as artistic activities (Badura et al., 2016; Darling, 2005; Darling et al., 2005; Eccles et al., 2003; Hansen et al., 2010; Wilson et al., 2010). However, this relationship was not significant in the study by Im et al. (2016). Linver et al. (2009) showed that students who participated only in sports EA had more positive results in academic achievement than those with little or no participation in organized activities, but worse academic achievement than those who participate in sports plus other activities, such as school, community, or religious groups.

We did not obtain a significant association between artistic EA and academic achievement, contrary to the results obtained in other studies, where artistic activities, not sports, provide greater reinforcement for academic achievement (Im et al., 2016), but only in girls (Luthar et al., 2006) or in late adolescents (Eccles et al., 2003).

Moreover, our work confirms the results of previous research by finding an association between reading books and academic achievement in early adolescence, namely, schoolchildren between 12 and 14 years old, the same age segment that Kingdon et al. (2017) reports. However, unlike our results, other research has found the same relation in late adolescence (e.g. Horbec, 2012; Whitten et al., 2019).

Regarding the second and third hypotheses (duration and breadth are associated with academic achievement), our results are in line with previous research, in which breadth (Fredricks, 2012; Fredricks and Eccles, 2006; Linver et al., 2009; Mahoney et al., 2006) and Duration (Darling, 2005; Darling et al., 2005; Eccles et al., 2003; Fredricks and Eccles, 2006; Zaff et al., 2003) relate to improvement of academic achievement. Nevertheless, our results show that duration is better predictor than breadth, contrary to previous literature (Barber et al., 2005; Eisman et al., 2016; Fredricks and Eccles, 2006).

However, we found that duration of EA is a predictor of academic achievement only in early adolescence, as Roth et al. (2010) and Eisman et al. (2016) show. On the other hand, in late adolescence, breadth is an important factor that enhances academic achievement as Eisman et al. (2016) show, but contrary to the results of Roth et al. (2010). Our findings confirm the evolutionary hypothesis that Eisman et al. (2016) present. In this sense, the stage of development of the young person would indicate the importance of the type of engagement (behavioral or psychological). Thus, in early adolescence, breadth would be more important, as those young people are more likely to seek a wide range of EA participation experiences (Busseri and Rose-Krasnor, 2009), to test which are more deserving of the investment of their free time. Thus, in late adolescence, they have already acquired greater control over time itself (Fredricks and Eccles, 2010), allowing more intense on specific activities and reducing the range of breadth (Busseri et al., 2006; Denault and Poulin, 2009b).

While the literature reports that participating in EA has a positive relationship with academic achievement, this is only true to a certain extent, since too long a duration has a negative impact (Fredricks, 2012; Fredricks and Eccles, 2010; Palma et al., 2014). Indeed, the overprogramming hypothesis (OSH) suggests a point at which the resulting participation becomes counterproductive in terms of benefits to the young person (Fredricks and Eccles, 2010; Mahoney et al., 2006). An explanation for this could be that much time invested in EA interferes with time that could be devoted to the family, curricular tasks, or other moments that the young person needs (Fredricks and Eccles, 2010; Mahoney and Vest, 2012; Palma et al., 2014), regardless of whether this occurs in early, middle, or late adolescence.

After all, the positive effects of participation in EA are manifest. Mahoney et al. (2005) reveal that participation in activities—voluntary, school and extracurricular—increases participation and school achievement because it facilitates the acquisition of interpersonal skills and positive social norms, membership in prosocial peer groups, and strengthening of emotional and social connections with the school itself.

Both the curricular and extracurricular schedules can trigger differential benefits in young people, nourishing the student in a beneficial way throughout adolescent development, through the interrelation of both contexts.

Limitations and future directions

According to Eisman et al. (2016), limitations of research on participation in EA and association with academic achievement may include a selection bias toward adolescents who start high school with higher levels of self-acceptance. They are more likely to participate (or be selected) in organized activities due to the high levels of skill required to participate, together with higher academic achievement.

Another limitation is that data on the intensity of participation have not been collected. It would be interesting to check in future studies how differences in intensity of participation over time relate to academic achievement by determining the proportion of time when participation occurs, as Vandell et al. (2015) shows. In any case, the main limitations of the investigation are the instrument used is a self-report and that it is a cross-sectional study, so longitudinal studies are required to make a causal explanation possible.

In fact, it should be added that the quantitative design prevents drawing broader conclusions. A multi-method design are more consistent and would enrich the information around the students’ perceptions, especially regarding the reasons why the activities are chosen. And, in this sense, would also deepen into the reasons why these activities are not chosen, as well as whether the offer suit all groups of adolescents.

An additional limitation is that no more complex models were evaluated (i.e., models with interactions). However, an attempt to correct that limitation included doing independent analyses for each of the age groups, but other factors can interact, such as sex with the other variables. But for statistical robustness, we do not consider that model. A longitudinal study that asks about behavior in recent years could try to correct that bias.

Besides, future research should also be directed to investigate the level of participation of students in these activities—performance, persistence, concentration, autonomy, etc.—and its relationship with their characteristics as students and their academic and social success in school.

In addition, this model is proposed for students with a common academic background. Therefore, students whose academic trajectories were not in the courses of the ordinary curricular approach were taken out of the study. We refer to students enrolled in curricular diversification or alternative curricular proposals for being unable to follow the ordinary academic trajectory.